1. Early intervention and social mobility: are pro-breastfeeding policies worth it? British Academy October 12 th 2011

2. The success story of the UNICEF Baby Friendly Initiative Emilia Del Bono and Birgitta Rabe October 12 th 2011

3. Cross-country differences in breastfeeding Source: Cattaneo et al. 2005, data for 1998-2001

4. Cross-country differences in breastfeeding Source: Cattaneo et al. 2005, data for 1998-2001

5. Breastfeeding in the UK: trends Source: UK Infant Feeding Survey, various years

6. Breastfeeding in the UK: mother’s education Source: UK Infant Feeding Survey, various years

7. Breastfeeding support

9. The Baby Friendly Hospital: Ten Steps to Successful Breastfeeding 1. Have a written breastfeeding policy that is routinely communicated to all healthcare staff 2. Train all healthcare staff in the skills necessary to implement the breastfeeding policy 3. Inform all pregnant women about the benefits and management of breastfeeding 4. Help mothers initiate breastfeeding soon after birth 5. Show mothers how to breastfeed and how to maintain lactation even if they are separated from their babies 6. Give newborn infants no food or drink other than breast milk, unless medically indicated 7. Practice rooming-in, allowing mothers and infants to remain together 24 hours a day 8. Encourage breastfeeding on demand 9. Give no artificial teats or dummies to breastfeeding infants 10. Foster the establishment of breastfeeding support groups and refer mothers to them on discharge from the hospital or clinic

10. The Baby Friendly Hospital: Ten Steps to Successful Breastfeeding 1. Have a written breastfeeding policy that is routinely communicated to all healthcare staff 2. Train all healthcare staff in the skills necessary to implement the breastfeeding policy 3. Inform all pregnant women about the benefits and management of breastfeeding 4. Help mothers initiate breastfeeding soon after birth 5. Show mothers how to breastfeed and how to maintain lactation even if they are separated from their babies 6. Give newborn infants no food or drink other than breast milk, unless medically indicated 7. Practice rooming-in, allowing mothers and infants to remain together 24 hours a day 8. Encourage breastfeeding on demand 9. Give no artificial teats or dummies to breastfeeding infants 10. Foster the establishment of breastfeeding support groups and refer mothers to them on discharge from the hospital or clinic Certificate of Commitment Full Accreditation

11. Kramer et al (2001) › Promotion of Breastfeeding Intervention Trial (PROBIT) in the Republic of Belarus › Long and very long term effects Cattaneo and Buzzetti (2001) › Implement a training programme based on BFI in Italy › Short and long term effects Dulon et al. (2003) › Evaluation of breastfeeding promotion in Germany › Long term effects Broadfoot et al. (2005) › Evaluation of UNICEF BFI initiative in Scotland › Short term effects Merten et al. (2005) › Evaluation of UNICEF BFI initiative in Switzerland › Long term effects Bartington et al. (2006) › Evaluation of UNICEF BFI initiative in the UK › Short term effects (initiation), but no long term effects Previous literature

12. Millennium Cohort Study + UNICEF BFI + National Sentinel Caesarean Section Audit Our data: 2000-2001

13. Effectiveness of BFI CC

14. Effectiveness of BFI Bf initiationExclusive bf at 4 wks

15. Effectiveness of BFI by subgroup

16. Contrary to previous findings, we find that the BFI increased breastfeeding rates not only at initiation, but also at 1 month Mothers giving birth in hospitals where BFI policies are fully implemented are 14.6 percent more likely to initiate breastfeeding and 6.6 percent more likely to continue to breastfeed exclusively at four weeks of age, in comparison to similar mothers in other hospitals No significant change in breastfeeding behaviour at longer durations however Policy most effective on less educated mothers and less advantaged groups Conclusions

17. Breastfeeding and infant hospitalisations for infections Amanda Sacker October 12 th 2011

18. In developing countries, exclusive breastfeeding has a large protective effect on infant mortality and severe morbidity  Is breastfeeding important for healthy, term infants in developed countries? We measured the effect of breastfeeding on hospitalization for diarrhoeal and lower respiratory tract infections (LRTI)  Largest study in a developed country  Any and exclusive current breastfeeding  Effects of stopping breastfeeding and starting solids  Account taken of a large range of other factors Introduction

19. % of mothers breastfeeding in first 8 months

20. Exclusive breastfeeding

21. Hospital admissions by 8 months (%)

22. Diarrhoea  18/10,000 babies who were not breastfed were hospitalised each month compared with 8/10,000 being partially and 5/10,000 exclusively breastfed  After taking account of other factors, exclusively breastfed babies 2.7 times less likely to be hospitalised Lower respiratory tract infection  49/10,000 babies who are not breastfed hospitalised compared with 25/10,000 partially and 30/10,000 exclusively breastfed  After taking account of other factors, babies were 50% less likely to be hospitalised if given any breast milk Breastfeeding and hospitalisation per month in the first 8 months after birth

23. Proportion of hospitalisations that could be prevented by breastfeeding

24. Months 1 – 4  Doubling of risk of being hospitalised for diarrhoea for every month since stopping breastfeeding  50 percent more likely to be hospitalised for LRTI for each month since stopping breastfeeding Months 5 – 7  30 percent more likely to be hospitalised for diarrhoea and 12 percent more likely for LRTI for each month since cessation of breastfeeding For diarrhoea, protective effect only persists for 1 month. Stops immediately on cessation for LRTI. How long do the effects last after stopping breastfeeding?

25. Could the explanation be that children who are not breastfed tend to go onto solids earlier Is there a beneficial effect of breastfeeding independent of feeding solids? Any alternative explanations?

26. Independent of feeding babies with solids, babies on formula were  2.5 times more likely to be hospitalised for diarrhoea than breastfed babies  1.5 times more likely to be hospitalised for a lower respiratory tract infection than breastfed babies Type of milk and feeding solids

27. Breastfeeding was associated with a lower risk of hospitalisation for diarrhoea and lower respiratory tract infection  Exclusive breastfeeding more beneficial than any breastfeeding  Benefits wore off very quickly after stopping breastfeeding  It doesn’t seem to be due to consuming solids or starting solids too soon Breastfeeding is beneficial for infant health even in developed countries Summary

28. Breastfeeding and socioemotional development Yvonne Kelly October 12 th 2011

29. Questions Is breastfeeding linked to reduced risk of socioemotional difficulties in young children? Do duration and exclusivity of breastfeeding matter?

30. Duration of breastfeeding

31. Prevalence of clinically relevant socioemotional difficulties

32. Likelihood of clinically relevant socioemotional difficulties by duration of any breastfeeding in term babies

33. Likelihood of clinically relevant socioemotional difficulties by duration of exclusive breastfeeding in term babies

34. Likelihood of clinically relevant socioemotional difficulties by duration of breastfeeding in preterm babies Any breastfeedingExclusive breastfeeding

35. Odds of emotional difficulties and hyperactivity by duration of exclusive breastfeeding in term babies Emotional difficulties Hyperactivity

36. Summary Breastfeeding appears to be linked to reduced risk of socioemotional difficulties Especially longer duration e.g. 4 months or more relates to 40% reduced likelihood of socioemotional difficulties The relationship remains regardless of socioeconomic and family contextual factors Policy needs to pay attention to longevity of breastfeeding

37. Breastfeeding and cognitive development Maria Iacovou October 12 th 2011

38. Breastfeeding and English SATS result at age 14 Breastfeeding (days) English SATs result (% of standard deviation)

39. Maternal education +++ Maternal social class The problem in a nutshell: causality Breastfeeding Child outcomes Housing +++

40. Randomised trials > Kramer et al 2008: 6-point difference in IQ scores Control for family background and maternal characteristics > Quigley, Hockley, Carson, Kelly, Renfrew and Sacker, 2011: Dose-response relationship; b/f 4m+ associated with several months development Sibling pairs > Evenhouse & Reilly 2005: small effect; Der et al 2006: negligible effect; Rees & Sabia 2009: significant effect Consider non-Western cultures > Brion et al 2011: Brazilian sample: ca. 6 IQ points; UK sample: ca. 3 IQ points. Getting round the problem of causality

41. Propensity score matching  Iacovou and Sevilla-Sanz 2010  Uses data from ALSPAC (Avon Longitudinal Study of Parents & Children) Instrumental variables  Del Bono and Rabe 2011  Uses data from Millennium Cohort Study Two papers from this project

42. BreastfedNot breastfed Outcome for breastfed babies Outcome for non-breastfed babies Born If babies randomly assigned, difference between these is “effect” of breastfeeding Propensity score matching in principle

43. BreastfedNot Breastfed Outcome for non-breastfed babies Born “Counterfactual”: outcome for non- breastfed babies IF they had been breastfed This counterfactual is not observed – so we synthesise it Difference between these represents “effect” of breastfeeding

44.                              The matching exercise…… Breastfed for 4 weeks or moreNot breastfed at 4 weeks Outcome for non-breastfed babies Outcome for breastfed babies who look JUST LIKE the non-breastfed babies

45. Results First, results – adjusted for family background but not “matched” Age 5 Age 7 Age 11 Age 14 Then, results from the matched samples

46. Maternal characteristics +++ Instrumental Variables in principle Breastfeeding Child outcomes +++ Instrument al Variable The Baby Friendly Initiative

47. Is BFI correlated with breastfeeding?  Yes (shown in previous presentation) Is BFI uncorrelated with maternal characteristics?  Yes: mothers don’t generally choose hospitals (subject of 2002/3 select committee report lamenting lack of information and birth choice) Instrumental variables in practice

48. Results Adjusted for family background, but not “instrumented” Age 5 Age 7 Age 3

49. IV results Much larger effects, BUT not precisely estimated Age 7 Age 5 Age 3

50. PSM analysis: effects of up to 10-15% of a standard deviation  Translates roughly to about 3 positions in a class of 30 children  Or, to 2-3 IQ points IV analysis: much larger effects  not statistically different from “naïve” estimates Generally smaller than some other studies have found BUT, we are looking at the effects of only 4 weeks’ breastfeeding Some evidence that these effects are larger at older ages How big are these effects?

51. Results differ. Why? PARTLY due to different data sets PARTLY due to different research designs  effects of initiation different to effects of breastfeeding for 4 weeks or 4 months  Also depends on the outcome analysed, and when it’s measured ALSO due to different analytical methods  Good statistical reasons why you would expect different techniques to produce different estimates – they are often measuring subtly different aspects of the relationship.  No “perfect” technique exists Making sense of it all

52. Breastfeeding does appear to be causally linked to better outcomes:  cognitive, health, behaviour Evidence of a dose/response relationship  more effect the longer a mother breastfeeds The Baby Friendly Initiative is effective  increases breastfeeding initiation BUT crucial to follow up BFI with other interve ntions  Longer durations are more beneficial than initiation alone  Initiation is necessary, but not sufficient, to achieve these longer durations The bottom line

53. Early intervention and social mobility: are pro-breastfeeding policies worth it? British Academy October 12 th 2011