Reasons for failure of breast-feeding counselling:mothers' perspectives in Bangladesh*R. Haider,1 1. Kabir,2 J.D. Hamadani,3 & D. Habte4

During the hospitalization in the Dhaka hospital of the Intemational Centre for Diarrhoeal Disease Research,Bangladesh, of a group of partially breast-fed infants aged 1-12 weeks who had been admitted with acutediarrhoea, their mothers were individually counselled by breast-feeding counsellors to start exclusive breast-feeding. The counselling was repeated 1 week later at home, and the women's infant-feeding practices wereevaluated 2 weeks after their infants had been discharged from hospital.

A total of25% of the mothers failed to breast-feed exclusively despite having been counselled. The casestudies of these mothers illustrate that although they generally complained about having "insufficient breastmilk'; various factors such as domineering grandmothers, lack of financial support by their husbands, toomuch housework, or disinterest contributed to their failure to breast-feed exclusively.

While family support is essential for all lactating mothers, women with familial or financial problemsrequire special attention and extra counselling sessions so that they can be helped to identify how to achieveand sustain exclusive breast-feeding.

IntroductionCurrently it is recommended by international healthagencies that all infants be exclusively breast-fed forat least 4 months, and if possible 6 months, and thatall mothers should be helped to breast-feed exclu-sively (1). In Bangladesh, a country usually noted forprolonged breast-feeding, the practice of exclusivebreast-feeding is not ingrained in the culture (2).Despite a claim in a national report of a dramaticincrease in exclusive breast-feeding (3), ongoingresearch findings fail to confirm this (R. Haider,unpublished data 1996, and El-Arifeen, personalcommunications, 1996).

The determinants of breast-feeding have beendescribed (4-6), but have not been disaggregatedby the type of feeding- exclusive and partial breast-feeding. It is important to identify these determi-nants in order to design specific strategies topromote exclusive breast-feeding (7, 8).

Promotion of breast-feeding has focused onmothers who deliver in hospitals and health facili-ties (9) but, in many developing countries, such as

* From: International Centre for Diarrhoeal Disease Research,Bangladesh (ICDDR, B), GPO Box 128, Dhaka 1000, Bangladesh.I Associate Scientist, Clinical Sciences Division, ICDDR, B GPOBox 128, Dhaka, Bangladesh. Requests for reprints should be sentto Dr Haider at this address.2 Scientist, Clinical Sciences Division, ICDDR, B.3 Senior Medical Officer, Clinical Sciences Division, ICDDR, B.4 Director, ICDDR, B.Reprint No. 5769

Bangladesh, only a small proportion of deliveriestake place in these institutions. Infants are usuallybrought to health facilities when they are ill. Sinceduring these visits there is an opportunity to influ-ence mothers on bre&t-feeding practices, we carriedout a study to provide and evaluate counselling formothers of young infants with diarrhoea who wereattending the Dhaka hospital of the InternationalCentre for Diarrhoeal Disease Research, Bangla-desh (ICDDR, B). The methods and results of thisstudy have been reported presiously (10). Briefly,60% of infants were being exclusively breast-fed and30%, predominantly breast-fed at the time of dis-charge from hospital. After 2 weeks at home, 75% ofthe mothers were breast-feeding exclusively. Thisarticle identifies some of the factors that might haveprevented the remaining 25% of mothers fromachieving or continuing exclusive breast-feeding, de-spite having been counselled during their hospitalstay.

Materials and methodsThe following WHO definitions for infants wereused (11):- exclusively breast-fed: if given only breast milk

(no other liquid or solid);- predominantly breast-fed: if given breast milk

plus water and/or oral rehydration salts solution(ORS); and

- partially breast-fed: if given other milk or gruel inaddition to breast milk.

Bulletin of the World Health Organization, 1997, 75 (3): 191-196 © World Health Organization 1997 191

R. Haider et al.

Two breast-feeding counsellors and two researchphysicians (all female) were trained for 3 weeks, us-ing the WHO/UNICEF breast-feeding counsellingtraining course (12). They were also taught aboutdiarrhoeal disease management, how to make theanthropometric measurements required for thestudy, and how to compare them with the NationalCenter for Health Statistics (NCHS) standards (13).

The selection criteria for the study, details ofcase management, counselling, and follow-up havealso been described previously (10).

Statistical methodsData were analysed using the SPSS/PC+ softwarepackage. The baseline data in the groups of motherswho were exclusively breast-feeding at the end of 2weeks, and those who were not, were compared. Alogistic regression analysis was performed on thosevariables that might affect the feeding status of theinfant, including the infant's age and sex, the moth-er's age and education level, and the father's income.

ResultsA total of 125 mother-infant pairs were recruited forbreast-feeding counselling. On follow-up at home 2weeks later, 21 mother-infant pairs could not beevaluated since 19 families had migrated out of thearea and two infants had died. There were no sig-nificant sociodemographic and feeding differencesbetween the mother-infant pairs who were lostto follow-up and those who completed the study.Among the remaining 104 mother-infant pairs, 80(77%) of the mothers were breast-feeding exclu-sively (two infants who received occasional waterhave been included with the exclusive group), whenthe infant's diarrhoea stopped and ORS was discon-tinued at home. A total of 24 mothers failed tobreast-feed exclusively and either continued or re-started other milk at home despite about 2 hours ofintensive counselling in the hospital.

The admission characteristics of the infants whowere exclusively breast-fed at 2 weeks were compa-rable with those who continued to be only partiallybreast-fed (Table 1). Table 2 shows the demographicfeatures of the infants' mothers in the two groups. Inboth groups the age, parity, and mean duration ofeducation of the mothers were comparable. Thehusbands of women who failed to breast-feed ex-clusively tended to have higher incomes than thehusbands of women who did, but the difference wasnot statistically significant.

There were no differences in the rates of exclu-sive breast-feeding among mothers who had re-

Table 1: Infant characteristics on admission to ICDDR,B hospital, Dhaka, by breast-feeding status at home 2weeks after discharge

Exclusively Partiallybreast-fed breast-fed(n= 80) (n= 24)

Age when CFa started (days) 17.2 ± 17.6b 17.12 + 20.3bAge at intervention (days) 52.6 ± 20.3b 47.0 ± 1 8.9bSex of babyFemale (n) 25 (31)c 8 (33)Male (n) 55 (69) 16 (67)

Baby bom at:Home (n) 59 (74) 17 (71)Hospital/clinic (n) 21 (26) 7 (29)

Duration of diarrhoea before 3.5 ± 1.7b 3.4 ± 1.4badmission (days)

Weight-for-age (% of NCHS 76.9 ± 11.3b 75.2 ± 1 1.0bmedian)

Weight-for-length (% of NCHS 91.2 ± 10.7b 87.8 ± 11.7bmedian)

Hospital stay (days) 4.5 ± 1.9b 3.9 ± 1.5b

a CF = complementary foods (other milk, gruel, etc.).b Mean + standard deviation.c Figures in parentheses are percentages.

ceived either two or three counselling sessions dur-ing their hospital stay. When mothers in the groupwho did not breast-feed exclusively were asked dur-ing evaluation at home why they gave additionalmilk to their babies, most of them complained ofhaving insufficient breast milk. Nevertheless, whenthe cases were reviewed in greater detail, includingthe breast-feeding counsellors' findings and impres-sions on the first week follow-up visit, other reasonsfor supplementation emerged, including the follow-ing: domineering grandmothers who advised themother that the infant be given other milk; anxiety

Table 2: Maternal characteristics and breast-feedingstatus of the study population on follow-up

Exclusive Partialbreast-feeding breast-feeding

(n =80) (n =24)

Mother's age (years) 22.8 ± 4.2a 20.9 ± 3.2aParity

First baby 38 (47)b 14 (58)Second baby 20 (25) 4 (17)Third baby 15 (19) 6 (25)

Mother's education level 4.2 ± 4.3a 4.7 ± 4.4a(years)

BMIc 19.7 ± 3.0a 19.0 ± 3.0aHusband's mean incomed 3940 ± 4377a 4682 ± 3871a

a Mean + standard deviation.b Figures in parentheses are percentages.c Body mass index = weight (kg)/height2 (m2).d In taka: US$1.00 = 40 taka.

WHO Bulletin OMS. Vol 75 1997192

Breast-feeding counselling: reasons for failure in Bangladesh

owing to financial insecurity; and mothers being un-willing to breast-feed exclusively (Table 3). Somecase studies (pseudonyms used) are given below.

Case studies

Raflq. Rafiq was the second baby born to the familyconcerned. His mother had no formal schooling andhis father was a shoemaker, earning Tk 5000 permonth (US$125). Born at home Rafiq was fed honeyafter birth, "refused" to suckle at the breast, andhence was bottle-fed fresh cow's milk on the firstday. He was 42 days old when admitted to the hospi-tal with diarrhoea. His mother breast-fed him exclu-sively for 1 day at the hospital (total stay, 3 days) andcontinued to do so for the first week at home; how-ever, his maternal grandmother had considerable in-fluence over family matters and sent "gripe water" (abanned product in Bangladesh) for the baby whenthe mother returned to the village. Rafiq's motherhad a minor road accident the week after he wasdischarged from hospital, and his grandmother be-lieved that her daughter's resultant blood loss woulddecrease the amount of breast milk that she couldproduce. Other family members also supported thisview and hence Rafiq's mother started giving himinfant formula within the second week.

Asim. Asim was the first baby born to educated par-ents, both mother and father having had 14 years ofeducation. His father, employed in a private firm,earned Tk 15000 per month (US$375). Asim was fedpowdered milk (infant formula) from the day ofhis birth, apparently because "his mother's milk didnot come". Asim was delivered by Caesarean sectionand his mother was sedated for a few hours; thehospital staff did not help her to initiate breast-feeding, and hence her mother-in-law requested thedoctor to suggest a milk formula. Initially fed milk

Table 3: Reasons given by mothersfeeding exclusively

for not breast-

Reason No. of mothersa

Influenced by baby's grandmother 8 (33)bHusband's advice 2 (8)Doctor's advicec 2 (2)Husband not/partly supporting financially or 5 (21)unemployed

Had to start working outside the home 2 (8)Too much housework 3 (12)Not willing to breast-feed exclusively 5 (21)a More than one reason was given in some cases.b Figures in parentheses are percentages.c In one case when the mother was admitted to another hospitaland in another when the baby was admitted to a private clinic.

with a dropper, Asim was bottle-fed within a fewdays. He was admitted to the ICDDR, B Dhakahospital with a 4-days' history of invasive diarrhoeaat the age of 54 days. During the hospital stay, Asimwas observed to be breast-feeding continuously; hismother complained that he would not leave thebreast even though he was not suckling all the time,and he was discharged while partially breast-fed.During the first week, his mother managed to de-crease the frequency of other milk to one feed perday, but soon after abandoned any plans for exclu-sive breast-feeding, stating that she had no servants,had to do the housework herself, and thus had insuf-ficient time for exclusive breast-feeding.

Tauhid. Tauhid was the first baby born to his par-ents. His mother had no formal education but hisfather had 5 years' schooling and earned Tk 6000 permonth (US$150) as a car driver. Tauhid was born bynormal delivery in hospital. His mother stated thatbecause her nipples were flat, he could not suckle atfirst; she did not receive any advice or help from thehospital staff about this problem. After trying to feedTauhid expressed breast milk and sugar water byspoon on the first day, his mother also started bottle-feeding powdered milk the same day. During thetime Tauhid was hospitalized with diarrhoea, hecried often and his mother complained of havinginsufficient breast milk. Tauhid's mother seemedvery anxious and, although not mentioned in heroriginal history, she later disclosed that her husbandhad other children from a previous marriage, livedwith his other wife, and provided her only withminimal financial support. During her stay in thehospital, she breast-fed partially for the first few daysand managed to breast-feed exclusively by the timeof discharge; however, she did not seem fully con-vinced about her ability to continue exclusive breast-feeding. She started giving Tauhid reconstitutedpowdered milk 2 days after discharge, after which hedeveloped a second episode of diarrhoea, and wasagain brought to ICDDR, B. During the follow-upvisit at week 1, he was being fed a soya-based for-mula on another doctor's advice, and by week 2, hewas receiving cow's milk.

Bithi. The third child of her parents, Bithi, born athome, first received fresh cow's milk at 15 days ofage and was admitted to the hospital with diarrhoeaat 33 days of age. Her father worked abroad, sendinghome Tk 10000 per month (US$250). Bithi's motherand siblings lived with her father's family. Dis-charged on exclusive breast-feeding, Bithi was againstarted on cow's milk during the first week at home.Her mother liked to go out. Despite being informedthat she would be visited at home after 1 week, she

WHO Bulletin OMS. Vol 75 1997 193

R. Haider et al.

did not stay at home and had to be located. AlthoughBithi's mother complained of insufficient breastmilk, the breast-feeding counsellor was able toexpress enough from her after the baby had breast-fed to show her that she had plenty. Since the babywas sick, the breast-feeding counsellor brought themother and baby along to the hospital. Hospital-ization was advised for the baby, but the motherrefused. The mother-in-law also favoured givingother milk to the baby, and Bithi's mother seemedgenerally unconcerned about her children. A fewdays before the follow-up visit, the elder daughter,aged about 3 years, was found wandering aroundsome distance from her home and was brought backby a neighbour. This same child had also been lost afew days previously.

DiscussionThe study has certain clear limitations. The group ofmothers who failed to breast-feed exclusively wassmall and since they came from various areas ofDhaka and were admitted to the hospital at differenttimes, we were unable to identify all the factors re-sponsible for their behaviour. Also, the motherswere anxious because their babies were sick and mayhave found it difficult to pay attention to apparentlyunrelated advice about breast-feeding. Furthermore,the short hospital stay did not allow time for ad-equate counselling.

A previous study in rural Bangladesh reportedthat three factors were associated with differencesin breast-feeding patterns: place of residence (urban/rural), maternal education, and income/socioeco-nomic status (14). In developed countries, more edu-cation and higher social class of the parents areassociated with higher breast-feeding rates amongwhite populations (15, 16). In our study, husbands'incomes in the partially fed group seemed to behigher but the trend was not significant. Duringthe counselling sessions, some of the mothers whoseeducation level and socioeconomic status werehigher attempted to justify partial breast-feedingby saying that since they could afford to give theirbabies powdered milk prepared correctly with cleanboiled water in sterilized feeding bottles why shouldthey bother with exclusive breast-feeding?

Attitudes are allegedly more important deter-minants of infant feeding behaviours than demo-graphic characteristics (17), with positive attitudesbeing more important predictors of initiation ofbreast-feeding than knowledge about breast-feeding(18). All the mothers enrolled in our study knew thatbreast-feeding was important and accepted counsel-ling in the hospital and on home follow-up; their

attitudes were therefore apparently positive to startwith. None the less, at the follow-up visits at home itbecame apparent that some of the mothers were notenthusiastic about exclusive breast-feeding. In otherstudies one of the reasons mothers have given forrejecting breast-feeding is that it limits their freedomand social life (19-21); although they did not ex-plicitly state this, it is likely that some of the mothersin the present study also felt this way.

The earlier mothers begin to breast-feed, thelonger they continue to do so (22). Exclusive breast-feeding, however, is a different issue, and so far onlyone study has specifically enquired about mothers'intentions in this regard (23). In our study, there wasno significant difference in the ages of their babieswhen mothers asked for counselling on exclusivebreast-feeding; and this therefore does not explainwhy one group failed to breast-feed.

Another important reason for bottle-feeding isthe lack of support from a significant "other person"(24). Investigators of infant-feeding practices haveidentified the baby's father as either the most in-fluential person in decision-making about feedingmethod (25) or an important source of support (26-28). In our study, the father's absence or not provid-ing financial support contributed to breast-feedingfailure in five cases. Kin, friend, and neighbour net-works have a significant impact on decision-makingabout breast-feeding (29), and although the mothersin our study may not have correlated their inabilityto breast-feed exclusively to the grandmother'spresence, the lactation counsellors identified thisas a probable negative influence. Possibly, the grand-mother also wished to contribute towards the child'sfeeding as part of caring practice, and so encouragefeeding the baby additional milk/gruel.

Some of the mothers who did not breast-feedexclusively stated that too much housework was thereason; however, studies of women with infants lessthan 1 year of age show that this is pot valid (30).

Mothers who received breast-feeding counsel-ling and support from the time of their babies' birthhad significantly greater rates of breast-feeding ini-tiation, exclusivity, and duration of total breast-feeding than those who did not (31, 32). In ourstudy, the counselling probably came much too latefor some of the mothers and therefore could notchange their practices. Peer counsellors living in themother's neighbourhood could be more influentialin this respect. We have trained peer counsellors forthis purpose and studies are being carried out toevaluate their impact on infant feeding practices.

Inclusion of key family members in the counsel-ling sessions at some stage is extremely important ifmothers are to be enabled to breast-feed exclusively.Also, mothers with family and financial problems

WHO Bulletin OMS. Vol 75 1997194

Breast-feeding counselling: reasons for failure in Bangladesh

may need more counselling sessions at the hospitalor at home so that they can be helped to identify howto achieve and sustain exclusive breast-feeding.

AcknowledgementsThis research was supported by WHO (grants. GURES/CDD/31 7/RB/94.B/300 and GURES/CDD/31 7NC/94.B/300) and the International Centre for Diarrhoeal DiseaseResearch, Bangladesh (ICDDR, B). ICDDR, B is sup-ported by countries and agencies that share its concernfor the health problems of developing countries. Currentdonors include the following: the aid agencies of thegovernments of Australia, Bangladesh, Belgium, Canada,China, Germany, Japan, Netherlands, Norway, Republicof Korea, Saudi Arabia, Sweden, Switzerland, UnitedKingdom and USA; international organizations including,the Arab Gulf Fund, Asian Development Bank, EuropeanUnion, IAEA, UNDP, UNFPA, UNICEF, and WHO; privatefoundations, including the Child Health Foundation, FordFoundation, Population Council, Rockefeller Foundation,and the Sasakawa Foundation; and private organizationsincluding American Express Bank, Bayer AG, CARE,Family Health International, Helen Keller International,International Committee of the Red Cross, the JohnsHopkins University, Procter & Gamble, RAND Corpora-tion, Sandoz, and University of California at Davis.

We thank the breast-feeding counsellors TanzilaFaruque and Shahara Bano. Dr S. Laston and Dr M. Aliare thanked for their helpful suggestions on improving themanuscript.

ResumeRaisons de I'6chec du conseil en matiered'allaitement: le point de vue des mbresau BangladeshAu Bangladesh, pays ou les meres allaitent gen6-ralement assez longtemps, I'allaitement maternelexclusif n'a pas de racines culturelles. Comprenantqu'il etait peut-etre possible d'influencer les pra-tiques des meres en matiere d'allaitement lorsqueles nouveau-nes sont malades, nous avons mendune 6tude consistant a dispenser des conseils auxmeres de nourrissons frequentant l'h6pital deDhaka (Centre international de recherche surles Maladies diarrheiques du Bangladesh), puis aevaluer cette operation.

Pour les besoins de l'etude, on a recrut6 125couples mere-enfant au cours de I'hospitalisationde nourrissons partiellement nourris au sein, agesde I a 12 semaines et trait6s pour diarrh6e aigue.Des conseilleres en allaitement ont conseille indi-viduellement aux meres d'allaiter leur enfant ex-clusivement au sein; ces conseils ont ete repet6sune semaine plus tard une fois la mere rentr6e chez

elle, et le mode d'alimentation des nournissonsa ete evalue deux semaines apres la sortie del'hopital.

Au total, 104 couples mere-enfant ont pu etresuivis a domicile 2 semaines plus tard. On n'a pasobserve de difference significative entre les carac-teristiques sociodemographiques et les modesd'allaitement des couples mere-enfant qui avaientete suivis et ceux qui ne l'avaient pas e. A la suitede la disparition de la diarrhee et de l'arret dessolutions de r6hydratation orale administrees adomicile, 80 meres (77%) nourrissaient leur enfantexclusivement au sein (deux enfants a qui l'ondonnait occasionnellement de l'eau ont et inclusdans le groupe des enfants exclusivement nourrisau sein). Au total, 24 meres avaient cess6 d'allai-ter exclusivement au sein et soit continuaient, soitavaient recommence a utiliser d'autres preparationsa domicile, malgre les conseils relativement inten-sifs qui leur avaient ete prodigu6s a l'h6pital.

Les conditions d'hospitalisation des nourris-sons exclusivement nourris au sein a 2 semainesetaient comparables avec celles des enfants quin'etaient que partiellement nourris au sein. L'age etla parite des meres 6taient comparables dans lesdeux groupes. La duree moyenne de scolarit6 desmeres etait egalement semblable. Les conjointsdes femmes qui avaient cesse de nourrir exclusive-ment au sein avaient tendance a se situer dans ungroupe a revenus plus eleves, mais la differencen'6tait pas statistiquement significative.

On n'a pas observe de difference dans les tauxd'allaitement maternel exclusif chez les meres quiavaient suivi deux ou trois s6ances d'informationpendant leur sejour a l'hopital. Au total, 25% desmeres avaient cesse de nourrir leur enfant exclu-sivement au sein malgre les conseils regus. Lesetudes de cas portant sur ces meres montrent que,meme si elles se plaignaient gen6ralement de "nepas avoir assez de lait", d'autres facteurs - unegrand-mere dominatrice, le manque de moyensfinanciers du mari, trop de taches menageres oule manque d'interet pour l'allaitement exclusifavaient contribue a l'echec.

Si le soutien de la famille est essentiel pour unemere qui allaite, les femmes presentant desproblemes familiaux ou financiers ont besoin d'uneattention particuliere ou bien de s6ances d'informa-tion supplementaires afin que l'on puisse les aidera trouver des solutions pour nourrir leur enfantexclusivement au sein.

References1. WHO/UNICEF. Breasifeeding in the 1990s: review

and implications for a global strategy, based on

WHO Bulletin OMS. Vol 75 1997 195

R. Haider et al.

the technical meeting, 25-28 June 1990, Geneva.Unpublished document WHO/MCH/NUT/90.2 (avail-able upon request from Family and ReproductiveHealth, World Health Organization, 1211 Geneva 27,Switzerland).

2. Talukder MQ-K. Bangladesh campaign for protectionand promotion of breastfeeding. Bangladesh journalof child health, 1992, 16: 25-31.

3. Progress of nations, 1996. New York, UNICEF, 1996.4. Forman MR. Review of research on the factors asso-

ciated with choice and duration of infant feeding inless-developed countries. Pediatrics, 1984, 74: 667-694.

5. Popkin BM et al. Breast-feeding determinants in low-income countries. Medical anthropology, 1983, 7: 1-31.

6. Winikoff B, Castle M, Laukaran V. Breastfeedingand bottle feeding controversies in the developingworld: evidence from a study in four countries. Socialscience and medicine, 1989, 29: 859-868.

7. Adair LS, Popkin BM, Guilkey DK. The durationof breast-feeding: how is it affected by biological,socioeconomic, demographic, health sector andfood industry factors? Demography, 1993, 30: 63-80.

8. Perez-Escamilla R. Breast-feeding patterns in LatinAmerica and the Caribbean. Bulletin of the Pan Ameri-can Health Organization, 1993, 27: 32-42.

9. Protecting, promoting and supporting breast-feeding.The special role of maternity services: a Joint WHO!UNICEF Statement. Geneva, World Health Organiza-tion, 1989.

10. Haider R et al. Breast-feeding counselling in adiarrhoeal disease hospital. Bulletin of the WorldHealth Organization, 1996, 74: 173-179.

11. Indicators for assessing breast-feeding practices.Report of an informal meeting, 11-12 June 1991,Geneva, Switzerland. Unpublished document WHO/CDD/SER/91.14 (available upon request from Divi-sion of Child Health and Development, World HealthOrganization, 1211 Geneva 27, Switzerland).

12. Breast-feeding counselling: a training course. Unpub-lished document WHO/CDD/93.3-6 and UNICEF/NUT/93:1-4, 1993 (available upon request from Divi-sion of Child Health and Development, World HealthOrganization, 1211 Geneva 27, Switzerland).

13. NCHS growth curves for children, birth-18 years,United States. Hyattsville, MD, National Centre ForHealth Statistics, 1987 (Series 11, No. 165-DHEW,Publication No. (PHS)78-1650).

14. Huffman SL. Determinants of breastfeeding in devel-oping countries: overview and policy implications.Studies in family planning, 1995, 41: 527-536.

15. Ryan AS et al. Recent declines in breast-feedingin the United States, 1984 through 1989. Pediatrics,1991, 88: 719-727.

16. Jones RAD, Belsey EM. Breastfeeding in an innerLondon borough: a study of cultural factors. Socialscience and medicine, 1977, 11: 175-179.

17. Jones DA. Attitudes of breast-feeding mothers: asurvey of 649 mothers. Social science and medicine,1986, 23:1151-1156.

18. Black R et al. Infant feeding decisions among preg-nant women from a WIC population in Georgia. Jour-nal of the American Dietetic Association, 1990, 90:255-259.

19. Bacon CJ, Wylie JM. Mothers' attitudes to infantfeeding at Newcastle General Hospital in summer1975. British medical journal, 1976, 7: 308-309.

20. Gielen-Carlson A et al. Determinants of breast feed-ing in a rural WIC population. Journal of human lacta-tion, 1992,8:5-11.

21. Bevan ML et al. Factors influencing breast feedingin an urban WIC program. Journal of the AmericanDietetic Association, 1984, 84: 563-567.

22. Goodline LA, Fried PA. Infant feeding practices: pre-and postnatal factors affecting choice of method andthe duration of breast feeding. Canadian journal ofpublic health, 1984, 75: 439-444.

23. Perez-Escamilla R et al. Exclusive breast-feedingduration is associated with attitudinal, socioeconomicand biocultural determinants in three Latin Americancountries. Journal of nutrition, 1995, 125: 2972-2984.

24. Freed GL, Jones TM, Schanler RJ. Prenatal deter-mination of demographic and attitudinal factors re-garding feeding practice in an indigent population.American journal of perinatology, 1992, 9: 420-424.

25. Novotny R et al. Health of infant is the main reasonfor breast-feeding in a WIC population in Hawaii. Jour-nal of the American Dietetic Association, 1994, 94:293-297.

26. Dusdiekier LB et al. Investigation of a model forthe initiation of breastfeeding in primigravida women.Social science and medicine, 1985, 20: 695-703.

27. Matich JR, Sims LS. A comparison of social supportvariables between women who intend to breast orbottle feed. Social science and medicine, 1992, 34:919-927.

28. James D, Jackson R, Probart C. Factors associatedwith breastfeeding prevalence and duration amonginternational students. Joumal of the American Di-etetic Association, 1994, 94: 194-196.

29. Bryant CA. The impact of kin, friend and neighbornetworks on infant feeding practices: Cuban, PuertoRican and Anglo families in Florida. Social scienceand medicine, 1982, 16: 1757-1765.

30. Cohen RJ et al. Maternal activity budgets: feasibilityof exclusive breastfeeding for six months amongurban women in Honduras. Social science andmedicine, 1995, 41: 527-536.

31. Sciacca JP et al. Influences on breastfeeding bylower-income women: an incentive-based, partner-supported educational program. Joumal of the Ameri-can Dietetic Association, 1995, 95: 323-328.

32. Kistin N, Abramson R, Dublin P. Effect of peercounsellors on breastfeeding initiation, exclusivity,and duration among low-income urban women. Jour-nal of human lactation, 1994, 10: 11-16.

196 WHO Bulletin OMS. Vol 75 1997