How do they make sure feeding research findings are accurate?

[CuppaTeaJanice]

I know there has been a lot of research into the benefits of BF against FF.

By its nature though, surely the only type of research possible would be observational studies. It wouldn't be ethical to force a woman to breastfeed, or to not breastfeed, for research purposes.

So how do the researchers make sure their findings are significant? How do they eliminate all other variables to ensure they are recording only differences caused by the type of milk consumed by a baby?

Some examples of variables might be...a woman prepared to put in the time and commitment to BF might be more inclined to ensure their toddler eats a fresh, organic, balanced diet, which would also have an impact on long term health, or there might be differences in mother's ages, socioeconomic status etc of people more likely to BF or to FF.

Much of the research would be long term studies, and I should imagine that to have large numbers of participants, in order to minimise any variations, would be prohibitively expensive.

I hope I've managed to word this so as not to offend anyone (I know threads on this topic can get quite heated sometimes!) but I'm genuinely interested in how the researchers come up with such strong pro-BF evidence when they have such little control over the participants in their studies.

In theory if you control for 'socioeconomic status' variables then you control for a lot of the 'quality of maternal care' differences that you mention.

Not because researchers are elitist, but because on average the less secure/comfortable your social position, the more stressed you are, and the harder it is to provide best quality parenting (on average, plenty of individual exceptions).

I know that studies have been done just to separate out skin-to-skin and breastfeeding, and while both are valuable in newborn development, they seem to have separate beneficial effects (small effects, but statistically significant).

Keep in mind that some studies don't find that breastfeeding is that advantageous, but most do, so it's also the trend of evidence. If Lots and Lots of studies seem to show the same thing, and much more often than showing the opposite thing, then that suggests a strong link, although it can never be quantified exactly. And if a finding appears in many studies by different people on different populations, then hopefully different unknown biases (things not measured) cancel out each other out, in affect.

it's very difficult and fact is that in most cases they're forced to speculate.

The only way to have absolutely valid results would be to have the exact same babies born twice, once they'd be BFed and once FFed, make sure that NO other variable changes and then compare physical, mental and emotional health and whathaveyou for each baby separately. I mean against its previous self -you knpow, the first time it as born.

Because such a study is impossible all the results we read about can be taken with a pinch of salt. I'm not arguing here that BM is not superior ro FM, only that we can't exactly know its benefits long term (if there are any that it).

oh dear, I did about 20 typos there in one paragraph, sorry...

You're right that most studies on bf are observational, Janice, but some get round this limitation. There is a big study (the PROBIT study) that randomised mothers to receive bf support or not bf support - they were not told how to feed their babies, but bf was either supported with a special programme, or not. The ones who got the special support did bf for longer and so comparisons could be made of health outcomes.

With large studies, you can observe and control for variables, and indeed with something as socio-economically mediated as infant feeding you absolutely must control for these, and with big samples you can do this. For instance the UK Millennium cohort study had something like 15,000 subjects, and you can crunch all sorts of numbers from that, knowing you are looking at the feeding alone, and not the 'type' of family the baby comes from.

It's highly likely, given what we know of physiology and gut development, that what an infant receives as his sole source of nourishment in the first weeks of life does have a health impact - you can look at babies before any impact of a toddler diet has had an effect, or you can look at a cohort over years, and ask the family what their other eating habits are, and control for this (like the Dundee study does), trying, again, to ensure you are isolating the infant feeding factor.

Many studies are funded long term, and while they are costly, I suppose, the value of cohort studies to public health is seen as so high it is worth it.

Ask me more questions, if you like I take a keen interest in research and research methodology.

mummygirl, you say "Because such a study is impossible all the results we read about can be taken with a pinch of salt."

This is just not true.

Pinch of salt? So no one can know anything about anything?

Read a good book on how to read a paper, and on how research is carried out - there are several. Read about the scientific method. It's not difficult - I have no formal scientific background, but by reading I learnt about it.

You cannot, surely, mean that research into the effects of any human behaviour is to be taken with 'a pinch of salt' because of the impossibility of being born and studied twice?

same as smoking - you can't force a group of people to smoke. Yet no one argues not smoking isnt beneficial.

NO, not any behaviour, I happen to have a Doctorate in Psychology so I have spent all my adult life working on research of human behaviour and the likes, carrying out statistical analyses for a living when skint. And that's why I know with confidence that there's a time when researchers just have to admit that all they can do is their best, but that it might not be precisely accurate.

If you read an actual research paper you will find that in the end there's always a disclaimer....

but if you're testing the health of elderly people who smoked for 10 years when they were teenagers and then stopped, there's only so much you can do to eliminate other factors

If you want a bit of relatively easy explanation of how scientific papers allow for things like social factors (among a lot of other variables), Bad Science by Ben Goldacre (available on amazon I think) is pretty good - lots of discussion of scientific method and the cumulative effect of lots of studies that suggest the same thing. I don't think there's anything specifically BF related in the book, but it will really help you understand (and sometimes question!) headlines surrounding scientific papers.

Thanks everyone this is really interesting. Is there somewhere I can read about research without having to trawl through difficult scientific papers - eg short paragraphs explaining basic findings of BF/FF research?

I guess what I'm most interested in is any research based on different proportions of BF/Mixed/FF and whether benefits decrease proportionally from one end of the spectrum to the other. Also any research on failed breastfeeding due to poor supply. I don't see how you would be able to research that ethically.

Human beings are not machines. They operate in unpredictable ways. And as you and others point out there are both practical and ethical limits to the kind of research that can be done on small babies and their mothers. But to be fair almost all studies of 'lifestyle' behaviour, including adult food choices etc has to be observational. Very few experimental studies which can completely control for all variables are possible. And even though the psychs don't like to hear this, it is pretty impossible to for control for unpredictable effects in the lab too.

So you have to apply the best research standard that you can upon population/epidemiological studies. And that is observations carried out on a large scale, with a sound methodology, which are repeatable by other researchers and which over time (eg demonstrated in a systematic review) bring up similar results. It's the best you can do and it is still valid.

Amazing really, when you consider the limitations of research into infant feeding, that we're so cavalier about giving babies formulas containing novel ingredients, when we know sod all about their long term affects on health and development.

mummygirl with a doctorate in psychology you obviously know a lot about research - you cannot believe that observational research has to be taken with a pinch of salt....surely? There are scientific, ethical, valid and robust ways of conducting observational research that give sound, repeatable and solid conclusions.

Breastfeeding does not have 'benefits'. Breastfeeding is the physiological norm. Formula feeding is the intervention. FF has been described as a massive, uncontrolled experiment on infant health and nutrition.

I agree that investigating the health of elderly people who smoked as teenagers 50 years ago and comparing it to people who have never smoked is gonna be hard - though there are studies that do track effects of behaviours over years and draw conclusions, controlling for variables (the Harvard Nurses study is one - thousands upon thousands of subjects and thousands of questions and investigations.

Fortunately, the health of babies and children is not quite so hard to track, as we are looking at a shorter period.

Janice - you can read abstracts of infant feeding research at www.babyfriendly.org.uk You can google the papers published as a result of the Millennium Cohort, but I think this is at the baby friendly link, too...the paper published about 2-3 years ago in Pediatrics does indeed show a 'dose response' effect ie the more formula, the less healthy in terms of hospital admissions for gastro-enteritis and other illnesses. A big US study whose name I have forgotten also shows dose response.

There is work on poor supply - why would this not be done ethically? You can investigate this, by taking case studies of women who have experienced it. This is a valid way of researching it - you don't have to randomise women to have poor supply somehow induced, in order to study it.

Thanks tiktok - I will check out that link.

RE research on poor supply. What I meant was that to study this they would have to have a number of participants with poor supply, and to have some supplement BM with FM. This could easily be studied on an observational basis. But to get the full balanced picture they would also need some participants with poor supply to persevere with exclusive BF. At what point would the researchers/ethics committee/medical profession step in and ensure the baby was receiving adequate nutrition? And would this not either create a conflict of interests with researchers wanting to attain significant results, against caring for their young participants, or mean that the study had to be stopped before its natural conclusion for ethical reasons, in which case the results would largely be invalid?

Janice, I understand what you mean, now, thanks.

I have never seen a study like the one you describe, and see the same issues as you do in it, anyway.

In addition, getting enough mothers together with poor supply that could not be fixed in some way would be very difficult....most (not all) cases of poor supply can be turned round by feeding more often, so you would not want to avoid trying to fix the supply issue just to adhere to study protocols. Also, we really never know, or rather can never predict, a cut-off point below which a bf baby really must get more calories into him. Obvious cases exist, but there is a whole gray area, full of apparently healthy but undoubtedly slow-growing babies; these are the babies you are asking about, I think - do these babies remain best off on breastmilk alone, or should they have formula? I don't think we have the research to give answers and instead each situation has to be judged individually. You might give a different answer for babies of different ages, for instance.

Very interesting.

Tiktok, is there any objective way of measuring supply on its own? It seems to me that all the ways of working out if a baby's getting enough milk (eg monitoring weight gain or test weighing) might be measuring supply, but they're also measuring how well the baby is feeding.

Lack of a good measurement method on its own might scupper any study of poor supply.

So many women worry about poor supply, and so many cite poor supply as a reason for topping up or switching to formula - it seems like a really important thing to look into.

But would it be possible to do this with current measurement methods?

welliemum - supply is dynamic...driven by the baby, and not static, that measuring 'supply' would be like measuring the speed of a car ie all you would get would be a snapshot of how much milk had been taken by the baby/how many mph the car was travelling at that moment.

Cars have a maximum speed, however, and it's not clear that women have a maximum production. Certainly, there are cases of mothers feeding triplets and more exclusively.

It does seem that women vary in their capabilities of production, though, and some women seem more sensitive to early day problems having a knock on effect later.

The other issue is that supply does not matter, as long as it is enough for the baby concerned. So a single small, slow growing baby will literally need less than a pair of large, fast growing twins.

There are some studies which track expressing yields, and they show a huge variation - maybe connected with the individual's pumping/hand expressing efficacy, rather than 'supply' though.

There is no reason why anyone needs to study supply, or that we would need to know the supply 'performance' of an individual. And I can't think how you would do it. It's what gets into the baby that counts, anyway...

Oh yes, I see that it's dynamic, and especially, that it just needs to be enough, not measure up to a standard.

What I was getting at was a bit different: a study tracking women's supply over time, in a way that could allow milk transfer to be isolated from supply (because obviously they're feeding back to each other).

It would be interesting (to me anyway) to build up a picture of what happens during a growth spurt, what happens if you give a bottle of formula at night, what happens if the baby gets a cold, what happens if you schedule feed ... it would need to be pretty big to be able to show the variation between individuals too.

I think so much of breastfeeding (at the start) is a leap of faith - you can't see the milk going in, or judge how that amount might be changing. It would be nice, I think to have a picture to show new breastfeeders of how responsive the system is - I can't speak for anyone else but I think I'd have found that very reassuring in the early days.

And it would be interesting to try and pin down in a quantitative way what the range of effect is of actions like giving a bottle - it could be helpful for people making that sort of practical decision. Obviously would be a range rather than a rule with big individual variation.

Dunno if this is something approaching what you're seeking, welliemum:

pediatrics.aappublications.org/cgi/content/abstract/117/3/e387

It shows the amazing variability in fat content of bf, and the wide variability in volume taken.

These were healthy, exc bf babies under 6 mths old.

The volume of milk taken at any one feed ranged from 0g to 240 g.

This shows it doesn't matter - as long as the baby is ok, sometimes the amount of milk he takes equals 0g

OK, I'll stop waffling on tiktok ! That study you've linked to is so interesting. I was particularly struck by the huge variation in volume produced. Do demand formula feeders find the same?

The other thing was that 64% of the babies were feeding at least once at night at age 1-6 months which is at odds with the popular idea that babies should be sleeping through from early on.

Wow, what a fascinating study. Thank you for linking to that tiktok.

It is a revealing study, and one which is a fine counter to anyone who worries that with breastfeeding, 'you can't see what they're getting'

I mean, if some healthy babies take 0g at a feed (not every feed, one assumes ) and some take 240g (about 9 oz!!) then we really don't need to be that concerned, I think.

I would love to see them repeat the same study design with babies aged 6 to 12 months. Though you'd want diaries of solids consumption too which I suppose it would be more work for the participating mothers.