Breastfeeding by Aboriginal mothers in Perth.
Objective: To document the breastfeeding practices of Aboriginal mothers delivering in Perth.
Design and methods: A cohort of mothers was followed from the time of delivery for six months to obtain details of infant feeding practices.
Subjects: A total of 455 mothers delivered during the study period and were asked to participate. A total of 425 mothers completed the initial questionnaire.
Setting: The study was undertaken in six public hospitals in Perth, Western Australia.
Data analyses: The data were analysed using SPSS. Breastfeeding duration was calculated using Kaplan-Meier survival analysis.
Main outcome measures: Breastfeeding initiation and duration.
Results: The mean age of the Aboriginal mothers was 21.8 years (range 14-39 years, SD 5.32) and the average gestational age was 38.3 weeks. Almost 50% of the mothers in the study delivered by Caesarean section. At discharge from hospital 89.4% (CI 86.6-92.1) of mothers were breastfeeding, declining to 58.8% (CI 53.5-64.1) at six months. When compared with non-Aboriginal mothers, the Aboriginal breastfeeding rates were higher than the non-Aboriginal average breastfeeding rates, but lower than the highest socioeconomic group.
Conclusions: The breastfeeding rates of Aboriginal mothers are higher than for other Australians. This is despite the low maternal age and level of education and the high rates of low-birth-weight infants and Caesarean section among this population. The World Health Organization recommendation for infant feeding is exclusive breastfeeding until six months, but less than one-third of Aboriginal mothers achieved this recommendation.
Key words: Aboriginal women, Australia, breastfeeding.
Breastfeeding of infants is important for the nutrition, and short-term and long-term health of the population. (1-3) A major National Health and Medical Research Council (NHMRC) review of Aboriginal and Torres Strait Islander nutrition, included information on breastfeeding. (4) Although this report acknowledged the difficulties in comparing breastfeeding rates from different studies because of varying definitions, the limited information available suggested that breastfeeding rates in rural areas, where the proportion of Aboriginal people is greater, were higher than for other Australians. (4) As a part of the National Breastfeeding Strategy a review was commissioned by the government of Aboriginal breastfeeding and breastfeeding promotion programs. (5) This report provided a comprehensive review of breastfeeding rates as reported in the literature and an inventory of breastfeeding promotion programs, and reached similar conclusions to the NHMRC report.
In the past, breastfeeding in traditional Aboriginal communities, and indeed all communities, was essential for the survival of the infant and usually continued for several years. (6,7) Breastfeeding was almost universal, but where it was not possible for any reason, another female relative within the group would feed the infant. Some evidence on breastfeeding practice in precontact times and during the nineteenth century and early twentieth century is available from ethnographic reports. Sellen has collected information on the infant feeding practices of 131 ethnic groups from non-industrialised societies, including several Aboriginal societies. (8) His review concluded that breastfeeding was universal and continued for an average period of 29 months.
There are many small studies of breastfeeding in Aboriginal communities, but they are often anecdotal reports or are based on small sample sizes. There are relatively few larger studies of breastfeeding and these are mainly cross-sectional in nature. (4) Based on a number of studies in Western Australia (WA), Gracey and colleagues reported that breastfeeding rates declined in groups who lived closer to urban areas. (9) A study of Aboriginal mothers was undertaken in Perth in the early 1980s. (10) Although the sample size was relatively small (n = 127), this showed a trend towards reduced rates of breastfeeding by urban Aboriginal people. After delivery 82% of Aboriginal mothers initiated breastfeeding; but by three months this had declined to 50%. After one year 20% of these mothers were still breastfeeding.
About 15 years later in Melbourne, 98% of Aboriginal mothers began breastfeeding, but only 50% were doing so at three months and 32% at six months. (11) Between the early 1980s and the late 1990s there was a substantial shift in attitudes of the Australian community to breastfeeding and the different trends described in these two studies might reflect the changes in the total community.
A national survey in 1994 by the Australian Bureau of Statistics collected information on breastfeeding as part of a larger cross-sectional study of health and social factors in Aboriginal Australians. (12) The collection of breastfeeding information by a questionnaire that asks mothers to remember events several years previously makes this study difficult to compare with the cohort studies cited elsewhere. However, in this study the Australian Bureau of Statistics found that 71% of children aged 12 years and under had been breastfed as infants. Of these, 27% were breastfed for 12 months or longer. More than 70% of Aboriginal babies in the Northern Territory were breastfed for six months or more. This proportion is much higher than that reported for other states/territories. Aboriginal babies in rural areas were more likely to be breastfed for longer than six months than those in urban areas. These data were reviewed in the Report on Aboriginal Nutrition by the NHMRC. (4)
There have been few studies of breastfeeding and morbidity in Aboriginal infants. However, in one review of asthma in Australia, Peat states that one reason for high rates of allergic disease in Aboriginal children might be low rates of breastfeeding. (13)
An enquiry into the health of the Aboriginal population by the Commonwealth Parliament highlighted the importance of nutrition to Aboriginal health. (14) 'Nutrition and antenatal care are critical foundations to future health and well-being. If mothers are malnourished then their babies will be born underweight and prone to considerable health problems. Malnourished children experience learning difficulties, subsequent unemployment problems and considerable health problems in later life'. It is surprising given this level of understanding of the community burden of poor nutrition, that the report does not mention the importance of protecting and promoting breastfeeding as one of the keys to good nutrition.
Because of the importance of breastfeeding to the health of infants, and the relative lack of prospective breastfeeding studies, it was decided to undertake a study of the infant feeding patterns of Aboriginal women in Perth. The specific objective was to document the infant feeding practices of Aboriginal mothers delivering in Perth. The use of a cohort to study breastfeeding patterns rather than a cross-sectional study is more expensive and more time-consuming, but gives more accurate results. (15)
The Perth Aboriginal Breastfeeding Study (PABS) used methodology very similar to the Perth Infant Feeding Study (PIFS) that has previously been published. (16,17) Mothers who agreed to participate in the present study were interviewed in hospital and again by telephone or personal interview at 2, 6, 10,14, 18 and 24 weeks postpartum. The inclusion criteria for the present study were mothers who identified themselves as Aboriginal and who were resident in the Perth Metropolitan Area. The study was conducted during 2001 and 2002. In WA all people who enter hospitals are asked if they identify themselves as Aboriginal. Many of the Perth hospitals provide special support for Aboriginal clients, and they make extra efforts to classify the ethnicity of their patients. The present study relied on self-identification and it is possible that some Aboriginal mothers did not choose to identify themselves on admission. The six public hospitals in the Perth Metropolitan region, which have delivery suites, were visited on a regular basis, every day when needed. The hospital nursing staff were most helpful in notifying that births had occurred. All Aboriginal mothers who delivered during the survey were contacted and invited to participate in the present study.
The data from the present study (PABS) were compared with data obtained from the earlier PIFS. In the latter study mothers who delivered at two suburban public hospitals were contacted and asked to participate in the study Details of the sample and methodology have been previously published. (16,18) A cohort of 556 Perth women completed an initial questionnaire and subsequently were followed up by telephone interview at 2, 6, 10, 14, 18 and 24 weeks post-partum, or until they ceased to breastfeed. The sample was consecutive and unselected. Those mothers agreeing to participate completed a self-administered baseline questionnaire while in hospital to determine breastfeeding initiation rates.
The original PIFS baseline questionnaire was designed to identify feeding method while in hospital and to collect information on variables known, or suspected, to be associated with breastfeeding initiation and duration, including socio-demographic, biomedical and psychosocial factors, along with hospital practices. As the baseline questionnaire was to be self-administered, a number of steps were taken to ensure that it was easy to read, comprehend and complete. The draft questionnaire was reviewed and modified by an adult literacy expert and then pilot-tested on another group of 20 new mothers. The Aboriginal nursing staff at Derbal Yerrigan Health Service also assisted in reviewing the language in the questionnaire. Wherever possible the wording of the PABS questionnaires was kept as close as practical to the PIFS to allow for later comparisons. (15,19,20)
There were differences in the follow-up policies for PIFS and PABS. In the PIFS, follow up was discontinued when breastfeeding ceased, but in PABS an attempt was made to follow up every mother for six months, in order to obtain additional information on weaning practices.
Data were entered and analysed using the Statistical Package for Social Sciences (SPSS for Windows, version 11, SPSS, Chicago, IL, USA). Descriptive procedures were used to calculate the prevalence of demographic-variables. Breastfeeding duration was calculated using Kaplan-Meier survival analysis. This type of analysis is used because of the presence of censored data, which in the present study refers to data from those cases where breastfeeding continued beyond the end of the study or beyond the time at which the subject dropped out of the study.
The purpose of the present study was explained to each mother and signed informed consent was obtained from those women agreeing to participate. The confidentiality of the data was assured and the mothers were assured that they could withdraw from the study at any time without prejudice. The NHMRC principles of research with Aboriginal communities were followed. (21) The project received initial ethics clearance from the Human Ethics Committee of Curtin University, the Department of Health WA, Aboriginal Health Section and the ethics committees of the participating hospitals.
A total of 455 identified Aboriginal mothers gave birth during the period of the present study and 425 completed the baseline questionnaire, a participation rate of 93%. The major Perth maternity hospital contributed 316 mothers to the present study whereas five other hospitals provided another 115 mothers. In the original population there were 441 singleton births, 13 sets of twins and one set of triplets.
The initial questionnaire was self-administered in hospital and the follow-up questionnaires were administered by telephone or by home visits. Follow up was very difficult and time-consuming, with frequent changes of address. During the period of the present study almost 80% of the original telephone numbers were disconnected. Wherever possible new numbers were found or home visits were made. At six months information was obtained on 239 mothers. During the six months 103 mothers moved to a rural address. To assess the validity of the sample at six months a comparison of demographic characteristics was undertaken between those interviewed and those lost to follow up. There was no significant difference in the age of those followed up and those lost to follow up; mean age of women followed up was 22.1 years compared with 21.2 years for those lost to follow up. Comparisons were also made for the type of delivery choice of initial feeding method, recent employment, educational level, smoking history and marital status, and no differences were found.
A comparison was made between the demographic characteristics of the participants in the PIFS and the PABS. The mean age of the Aboriginal mothers was 21.8 years (range 14-39 years, SD 5.32) compared with the PIFS mothers 27.6 years (range 16-44 years, SD 5.44). The mean parity for Aboriginal mothers was 1.6 compared with 2.1 for the PIFS mothers. As can be seen from Table 1 there were important differences in most of the attributes measured.
The major differences between the two groups were in education level, marital status, partners' occupation and parity The PIFS was biased towards a lower socioeconomic sample by the choice of the hospitals, as those hospitals chosen were known to have a higher number of public admission patients and were located in lower socioeconomic areas, but even so the difference between the PIFS and the PABS highlighted the economic deprivation of the Aboriginal participants.
The median gestational age was 40 weeks for the PABS (17.6% <40 weeks, range 23-41) and 40 weeks for the PIFS (20.7% <40 weeks, range 34-42). The method of delivery is shown in Table 2 with the Aboriginal mothers having more than twice the Caesarean section rate of the PIFS study.
The duration of breastfeeding was examined using Kaplan-Meier survival analysis. There was a significant difference between the PABS and the PIFS; the log rank test statistic was 5.34, P = 0.021. The rates of breastfeeding at four weekly intervals are shown in Table 3.
The PIFS data were also analysed for breastfeeding rates by occupational status using the Congalton scale. A comparison with the PABS data is shown in Figure 1.
The median age of infants at which Aboriginal mothers introduced complementary foods was 3.9 months. The median time was similar for mothers who breastfed for six months and mothers who breastfed for a shorter period of time. The rate of 'any breastfeeding' at 24 weeks was 58%, but the rate of exclusive breastfeeding was only 31.5%. The most popular first solid foods to be introduced were infant cereal preparations with a median age of introduction of 4.7 months.
Breastfeeding initiation rates and rates to 24 weeks for Aboriginal mothers are higher than for non-Aboriginal mothers in Perth. This finding is consistent with the results of earlier national studies. The 1995 National Health Survey (NHS) showed that Aboriginal mothers breastfed for longer time than their non-Aboriginal counterparts. (12,22) Although this result is based on a small number of Aboriginal mothers and does not include mothers (Aboriginal or non-Aboriginal) from remote areas, it is consistent with the result of the 1994 National Aboriginal and Torres Strait Islander Survey. (23) In a cross-sectional study in Perth of 273 Aboriginal mothers in Perth, 89% initiated breastfeeding. (24)
When the PIFS mothers are classified by socioeconomic status, the upper class non-Aboriginal mothers have higher breastfeeding rates than the PABS mothers. An earlier study conducted in Perth reported that urban Aboriginal women breastfed for approximately the same length of time as non-Aboriginal women of low socioeconomic status. (10,25) One study found no difference in breastfeeding rates between Aboriginal and non-Aboriginal women of low socioeconomic backgrounds. (26) The PABS data show that two decades later, Aboriginal mothers are more successful in breastfeeding than non-Aboriginal mothers of lower socioeconomic status.
The exclusive breastfeeding rate for Aboriginal mothers at six months was 31.5%. Comparative data are not available from the PIFS, but the rate of fully breastfed infants at six months from the 1995 NHS was 18.6%. (27) The comparable figures for any breastfeeding at six months are 58.8% and 46%. Although caution needs to be exercised in directly comparing these studies as the NHS was a retrospective survey compared with the prospective cohort in the present study.
The PABS sample was characterised by the lower average age of the mothers, the lower levels of their education and the lower levels of their socioeconomic status. These are all factors that are well documented as being associated with lower rates of breastfeeding initiation and duration. There is a higher prevalence and a longer duration of breastfeeding by mothers from higher socioeconomic groups who are better educated, are older, and have previously breastfed. (20,28-32) This means the rates of breastfeeding achieved in the present study were achieved by mothers who normally would be regarded as being of high risk for low rates of breastfeeding.
[FIGURE 1 OMITTED]
There are some limitations that apply to the PABS. Aboriginal mothers were all self-identified. Mothers who chose not to identify themselves as Aboriginal would not be included. Practically this was not a major concern as the major hospital in the present study has several Aboriginal staff people to provide additional care to Aboriginal mothers. The follow-up rate at six months was 58% of those enrolled in the present study. There were no statistically significant differences found in the demographic characteristics between those women who completed the study and those who chose not to participate. Both the PIFS and the PABS only included public patients. Although there are very few Aboriginal births in private hospitals, about 38% of non-Aboriginal births in the metropolitan area occur in private hospitals. (33)
Breastfeeding rates of urban Aboriginal women and non-Aboriginal women are very similar for the first month after birth, but become different by six months. Breastfeeding initiation rates for the Aboriginal mothers are quite high at almost 90%. However, there is a decline to about 59% at 24 weeks, with only 31.5% being exclusively breastfed at this time. The current World Health Organization recommendation of exclusive breastfeeding to six months is obviously not being met.
What is impressive about the Aboriginal mothers is that they have achieved these breastfeeding rates, despite the high rates of Caesarean section, high numbers of infants needing admission to a special care nursery and the low maternal age. Although initiation rates are high, there is considerable scope to increase the duration of breastfeeding, and duration of exclusive breastfeeding needs to be improved. Increasing the percentage of breastfeeding at six months to 80% could be a realistic target in the medium term.
We gratefully acknowledge the willing assistance given by the mothers in the present study, the hospital staff and community health workers. Without this assistance the present study would not have been possible.
1 World Health Organization (WHO). Infant and young child nutrition. Global Strategy on Infant and Young Child Feeding. Report by the Secretariat Fifty-Fifth World Health Assembly. Provisional agenda item 13.10. 2002.
2 World Health Organization. Department of Nutrition for Health and Development. Report of the expert consultation on the optimal duration of exclusive breastfeeding. Geneva, Switzerland, 28-30 March 2001. 2002. Report No.: WHO/NHD/01.09.
3 Graham K, Scott JA, Binns CW, Oddy W. National targets for breastfeeding at hospital discharge have been achieved in Perth. Acta Paediatrica 2005; 94: 352-6.
4 NHMRC. Nutrition in Aboriginal and Torres Strait Islander Peoples: An Information Paper. Canberra: National Health and Medical Research Council, 2000.
5 Engeler T, McDonald M, Miller M, Groos A, Black M, Leonard D. Review of Current interventions and Identification of Best Practice Currently Used by Community Based Aboriginal and Torres Strait Islander Health Service Providers in Promoting and Supporting Breastfeeding and Appropriate Infant Nutrition. Canberra: Commonwealth of Australia, 1998.
6 Hitchcock N. Infant feeding in Australia: an historical perspective. Part 3: Australian Aborigines and recent migrants. Aust J Nutr Diet 1989; 46: 108-11.
7 Gracey M. Maternal health, breast-feeding and infant nutrition in Australian aborigines. Acta Paediatr Jpn 1989; 31: 377-80.
8 Sellen D. Comparison of infant feeding patterns reported for non-industrialised populations with current recommendations. J Nutr 2001; 131: 2707-15.
9 Gracey M, Murray H, Hitchcock N, Owles E, Murphy B. The nutrition of Australian Aboriginal infants and young children. Nutr Res 1983; 3: 133-47.
10 Phillips F, Dibley M. A longitudinal study of feeding patterns of Aboriginal infants living in Perth, 1980-1982. Proc Nutr Soc Aust 1983; 8: 130-32.
11 Holmes W, Phillips J, Thorpe L. Initiation rate and duration of breastfeeding in the Melbourne Aboriginal Community. Aust N Z J Public Health 1997; 21: 500-503.
12 Australian Bureau of Statistics (Australian Bureau of Statistics). The Health and Welfare of Australia's Aboriginal and Torres Strait Islander Peoples 1997. 1997. Report No.: 4704.0.
13 Peat J, Veale A. Impact and aetiology of respiratory infections, asthma and airway disease in Australian Aborigines. J Paediatr Child Health 2001; 37: 108-12.
14 House of Representatives. Health is Life: Report on the Inquiry into Indigenous Health. House of Representatives Standing Committee on Family and Community Affairs. Canberra, ACT: Commonwealth of Australia, 2000.
15 Scott J, Binns C, Aroni R. Breastfeeding Perth: recent trends. Aust N Z J Public Health 1996; 20: 210-11.
16 Scott JA, Binns CW, Aroni RA. The influence of reported paternal attitudes on the decision to breastfeed. J Paediatr Child Health 1997; 33: 305-7.
17 Binns CW, Scott J. Breastfeeding: reasons for starting, reasons for stopping and problems along the way. Breastfeed Rev 2002; 10: 13-19.
18 Scott JA, Aitkin I, Binns CW, Aroni RA. Factors associated with the duration of breastfeeding amongst women in Perth, Australia. Acta Paediatr Scand 1999; 88: 416-21.
19 Scott JA, Binns CW, Aroni RA. The influence of reported paternal attitudes on the decision to breast-feed. J Paediatr Child Health 1997; 33: 305-7.
20 Scott JA, Binns CW. Factors associated with the initiation and duration of breast feeding. Aust J Nutr Diet 1998; 55: 51-61.
21 NHMRC (Commonwealth of Australia). Guidelines on Ethical Matters in Aboriginal and Torres Strait Islander Health Research. Approved by the 111th Session of the National Health and Medical Research Council, Brisbane, June 1991. 1991.
22 ABS, AIHW (Australian Bureau of Statistics). The Health and Welfare of Australia's Aboriginal and Torres Strait Islander Peoples. 1999. Report No.: ABS Cat. No. 4704.0.
23 ABS (Australian Bureau of Statistics). National Aboriginal and Torres Strait Islander Survey 1994: Detailed Findings. 1995. Report No.: Catalogue No. 4190.0.
24 Eades S, Read A, Team BG. Infant care practices in a metropolitan Aboriginal population. J Paediatr Child Health 1999; 35: 541-4.
25 Hitchcock NE, Coy JF. Infant feeding practices in Western Australia and Tasmania: a joint survey, 1984-1985. Med J Aust 1988; 148: 114-17.
26 Cox JW. The antenatal and perinatal characteristics of socio-economically depressed Caucasians. Aust N Z J Obstet Gynaecol 1981; 21: 20-23.
27 Donath S, Amir L. Rates of breastfeeding in Australia by state and socio-economic status: evidence from the 1995 National Health Survey. J Paediatr Child Health 2000; 36: 164-8.
28 Hartmann PE. Lactation and reproduction in Western Australian women. J Reprod Med 1987; 32: 543-7.
29 Hitchcock NE, Coy JF. The growth of healthy Australian infants in relation to infant feeding and social group. Med J Aust 1989; 150: 306-11.
30 Kocturk T, Zetterstrom R. Breastfeeding and its promotion. Acta Paediatr Scand 1988: 183-90.
31 Milligan RA, Pugh LC, Bronner YL, Spatz DL, Brown LP. Breastfeeding duration among low income women. J Midwifery Womens Health 2000; 45: 246-52.
32 Simopoulos AP, Grave GD. Review of research on the factors associated with choice and duration of infant feeding practice. Pediatrics 1984; 74: S603-14.
33 Department of Health WA (DOH). Perinatal Statistics in Western Australia, 2003. Perth: Department of Health (WA), 2004.
Colin W. BINNS, (1) Dawn GILCHRIST, (2) Beth WOODS, (1) Michael GRACEY, (1) Jane SCOTT, (3) Hannah SMITH (nee Herod), (1) Min ZHANG (1) and Brian ROBERMAN (4)
(1) School of Public Health, Curtin University of Technology, Perth, (2) Goldfields South-East Health Region, Kalgoorlie, and (4) King Edward Memorial Hospital, Perth, Western Australia, Australia; and (3) Division of Developmental Medicine, University of Glasgow, Glasgow, UK
C.W. Binns, MB BS, MPH, Professor
D. Gilchrist, MPH, Manager
B. Woods, RN, Research Officer
M. Gracey, AO, MD, PhD, Professor
J. Scott, BAppSc, GradDipDiet, MPH, PhD, Senior Lecturer
H. Smith (nee Herod), MPH, Research Officer
M. Zhang, PhD, MD, GradDipHealthSci, Research Fellow
B. Roberman, MB BS, FRANZCOG, FRCOG, Senior
Correspondence: C.W Binns, Curtin University of Technology, School of Public Health, GPO Box U1987, Perth, WA 6845, Australia.
Table 1 Demographic characteristics of participants in the Perth Infant Feeding Study (PIFS) and the Perth Aboriginal Breastfeeding Study (PABS) (%) PABS % PABS follow up 24 PIFS % Characteristic (n = 425) weeks (n = 239) (n = 556) Highest level of education completed Before year 10 40.5 45.9 14.0 Year 10 44.5 37.6 38.5 Year 12 5.9 7.7 17.4 Trade/dip/TAFE 5.6 5.0 20.9 Bachelor degree/+ 2.4 2.0 6.3 Not stated 1.2 1.0 2.9 Marital status Never married 11.8 15.1 6.4 Now married 4.5 6.5 70.8 Defacto 82.6 77.8 21.4 Divorced or separated 0.5 0.5 1.4 Not stated 0.7 1.0 Partners occupation Professional/administrative 0.0 0.0 17.8 Sales/clerical/trades/labourers 1.9 2.3 64.4 Self-employed 0.0 0.0 3.8 CEPD 7.8 11.8 0.0 Unemployed 74.1 75.3 3.4 Other (a) 12.2 10.1 1.4 Not stated 4.0 3.0 9.2 (a) Includes students and invalid pensioners. CDEP, community development employment projects; TAFE, technical and further education. Table 2 Biomedical characteristics--comparison of initial and follow-up samples with the Perth Infant Feeding Study (PIFS) (non-Aboriginal) PABS Follow up 24 PIFS (n = 425) weeks (n = 239) (n = 556) Gestational age (%) <40 weeks 17.6 11.3 20.7 <35 weeks 14.0 8.1 0.7 Method of delivery (%) Spontaneous vaginal delivery 46.5 51.0 68.6 Assisted vaginal delivery with 2.6 3.7 13.6 forceps or suction Caesarean section 49.4 44.8 17.6 Born at home 0.2 0.4 0.0 Missing 0.4 0.0 0.0 Parity (%) 1 73.4 74.8 30.7 2 8.7 7.9 38.0 3 6.4 6.7 18.4 4 4.5 3.4 9.4 5 1.7 1.3 2.4 6+ 3.9 4.6 1.1 Missing 1.4 1.3 0.0 Infant birth weight (%) <2500 g 22.6 13.0 4.5 Infant admitted to special nursery (%) Yes 28.7 (a) 18.4 (a) 15.8 (a) The proportion of infants admitted to special nursery is significantly different (P = 0.0032, [chi square] = 8.6). PABS, Perth Aboriginal Breastfeeding Study. Table 3 Percentage of Perth Aboriginal women and non-Aboriginal women in the Perth Infant Feeding Study (PIFS) breast-feeding (a) at hospital discharge and selected time points PABS total PABS adjusted (b) PIFS total Interview % (95% CI) % (95% CI) % (95% CI) At discharge 89.4 (86.6-92.1) 89.0 (86.3-91.7) 83.8 (80.9-86.6) 4 weeks 75.1 (70.6-79.6) 75.0 (70.5-79.5) 74.8 (71.1-78.5) 8 weeks 69.0 (64.1-73.9) 69.0 (64.2-73.8) 68.5 (64.5-72.5) 12 weeks 66.2 (61.2-71.2) 65.8 (60.8-70.8) 62.0 (57.7-66.2) 16 weeks 63.8 (58.7-68.9) 63.2 (58.1-68.3) 58.0 (53.6-62.3) 20 weeks 58.8 (53.5-64.1) 57.9 (52.6-63.2) 52.5 (48.0-56.9) 24 weeks 58.8 (53.5-64.1) 57.9 (52.7-63.2) 49.9 (46.0-53.8) (a) 'Any breastfeeding'. (b) Adjusted for the slightly lower proportion of infants who were admitted to intensive care who were followed up. Sample size: Perth Aboriginal Breastfeeding Study (PABS) n = 425 declining to 239 at 24 weeks, PIFS n = 556 declining to 217. Note the sample size figures are not comparable because of differences in follow-up methods noted in the 'Methods' part.