ebook img

Determinants of breastfeeding indicators among children less than 24 months of age in Tanzania: a secondary analysis of the 2010 Tanzania Demographic and Health Survey. PDF

0.28 MB·English
by  VictorRose
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Determinants of breastfeeding indicators among children less than 24 months of age in Tanzania: a secondary analysis of the 2010 Tanzania Demographic and Health Survey.

Open Access Research Determinants of breastfeeding indicators among children less than 24 months of age in Tanzania: a secondary analysis of the 2010 Tanzania Demographic and Health Survey Rose Victor,1 Surinder K Baines,1 Kingsley E Agho,2 Michael J Dibley3 Tocite:VictorR,BainesSK, ABSTRACT ARTICLE SUMMARY AghoKE,etal.Determinants Objective:ToexaminetheprevalenceofkeyWHO ofbreastfeedingindicators breastfeedingindicatorsandidentifydeterminantsof Article focus amongchildrenlessthan 24monthsofagein suboptimalbreastfeedingpracticesamongchildren ▪ ThispaperaimstoexaminekeyWHObreastfeed- Tanzania:asecondary agedlessthan24monthsinTanzania. ing indicators in Tanzania and determine factors analysisofthe2010Tanzania Design, setting andparticipants:Secondary associated with delayed initiation of breastfeed- DemographicandHealth analysesofcross-sectionaldatafromthe2010 ing, non-exclusive breastfeeding and predomin- Survey.BMJOpen2013;3: TanzaniaDemographicandHealthSurvey.Thesurvey ant breastfeeding practices in children aged e001529.doi:10.1136/ usedastratifiedtwo-stageclustersampleof10312 0–23months. bmjopen-2012-001529 householdsfromeightgeographicalzonesofTanzania. Key messages Thesampleconsistedof3112childrenaged ▸ Prepublicationhistoryand 0–23months. ▪ Prevalence of early initiation and exclusive additionalmaterialforthis Main outcome measures:Outcomemeasureswere breastfeeding indicators fell below national paperareavailableonline.To factorssignificantlyassociatedwithdelayedinitiationof targets for Tanzania. A considerable proportion viewthesefilespleasevisit breastfeeding,non-exclusivebreastfeedingand of infants less than 6months were predomin- thejournalonline antlybreastfed. predominantbreastfeedinginthefirst6months. (http://dx.doi.org/10.1136/ ▪ Children who live in the Northern, Southern Results:Breastfeedingwasinitiatedwithinthefirst bmjopen-2012-001529). zones and Zanzibar were at higher risk of sub- hourofbirthin46.1%ofmothers.Ininfantsagedless optimal\optimalbreastfeedingpracticesthanchil- Received23May2012 than6months,theprevalenceofexclusive dreninothergeographicalzonesofTanzania. Revised6October2012 breastfeedingwas49.9%butonly22.9%were Accepted1November2012 exclusivelybreastfedat4–5months.Seventeenper ▪ Young maternal age, lower maternal education, employment, home delivery and lack of profes- centofinfants,lessthan6monthsofage,were Thisfinalarticleisavailable ‘predominantlybreastfed’.At12–15months,94.0% sionalassistance at birth were the maindetermi- foruseunderthetermsof nants of suboptimal breastfeeding practices in ofinfantswerestillbreastfedbuttheproportion theCreativeCommons decreasedto51.1%at20–23monthsofage. Tanzania. AttributionNon-Commercial 2.0Licence;see Multivariateanalysisrevealedthattheriskofdelayed Strengths and limitations of this study http://bmjopen.bmj.com initiationofbreastfeedingwithin1hafterbirthwas ▪ The main limitation was the cross-sectional significantlyhigheramongyoungmothersaged nature of the survey, which limited inferences <24years,uneducatedandemployedmothersfrom aboutcausalityfromtheanalyses. ruralareaswhodeliveredbycaesareansectionand ▪ In addition, exclusive breastfeeding was based thosewhodeliveredathomeandwereassistedby on a 24h recall rather than a longer recall traditionalbirthattendantsorrelatives.Theriskfactors period, and this short recall may have missed associatedwithnon-exclusivebreastfeeding,duringthe some infants who were fed other liquids or first6months,werelackofprofessionalassistanceat foodspriorto24hbeforethesurvey. birthandresidenceinurbanareas.Theriskof ▪ However, the use of a large nationally represen- predominantbreastfeedingwassignificantlyhigher tative survey sample with a very high response amonginfantsfromtheZanzibargeographicalzone. rate (96.4%), appropriate statistical adjustments Conclusions:Earlyinitiationofbreastfeedingand forsurveydesignandmodellingforconfounding Fornumberedaffiliationssee exclusivebreastfeedingindicatorswereunsatisfactory effects add strength to the validity of the find- endofarticle andarebelowthenationaltargetsforTanzania.To ings. Furthermore, restricting the sample to only improvebreastfeedingpractices,nationallevel children less than 2years who lived with their Correspondenceto programmeswillberequired,butwithafocusonthe mothers helped ensure greateraccuracyof infor- RoseVictor; targetgroupswithsuboptimalbreastfeedingpractices. mationregardingbreastfeedingpractices. [email protected]; [email protected] VictorR,BainesSK,AghoKE,etal.BMJOpen2013;3:e001529.doi:10.1136/bmjopen-2012-001529 1 Determinants of breastfeeding indicators in Tanzania INTRODUCTION differences.16 18 These factors have been documented to WHO1 infant-feeding guidelines recommend that all be either positively or negatively associated with breast- infants should be breastfed within 1h after birth and feeding practices, and the inconsistencies of the results exclusively breastfed from birth until 6months of life. foundindifferentcountriesmakeitdifficulttogeneralise Thereafter, infants should be introduced to nutritionally the findings to all countries, hence the need to identify adequate and safe complementary foods with continued factorsthat are associated with breastfeeding practices in breastfeeding for up to 2years or beyond. In line with Tanzania. the WHO recommendations, Tanzania has been imple- This secondary data analysis of the 2010 TDHS aims menting a number of initiatives to improve infant- to describe the prevalence of breastfeeding practices feeding practices, which include the National Strategy using the current WHO breastfeeding indicators,19 and and Implementation Plan on Infant and Young Child to determine the factors associated with delayed initi- Nutrition, the baby-friendly hospital initiatives and the ation of breastfeeding, non-exclusive breastfeeding and training of health workers on infant-feeding skills. predominant breastfeeding among children less than Despite these efforts, breastfeeding practices and espe- 24months of age inTanzania. cially early initiation and exclusive breastfeeding remain suboptimal in Tanzania.2 METHODS According to the Tanzania Demographic and Health Survey (TDHS) of 2010,2 breastfeeding was almost uni- Data source The present analysis was based on the 2010 Tanzania versal at 99% in all sociodemographic categories; Demographic and Health Survey (TDHS),2 which was however, early initiation of breastfeeding, within 1h conducted from December 2009 to May 2010 by the after birth, was reported by 46.1% of women who National Bureau of Statistics and the Office of the Chief recently delivered a baby, whereas the prevalence of Government Statistician –Zanzibar in collaboration with exclusive breastfeeding (EBF) was 50% among infants less than 6months.2 This implies that a considerable the Ministry of Health and Social Welfare. The 2010 TDHS is the eighth in a series of Demographic and proportion of infants aged less than 6months are intro- Health Surveys conducted in Tanzania. The survey duced toother liquids and solid foodsbeforethe recom- aimed to gather information about child mortality,nutri- mended age of 6months and thereby limit the full benefits of breastfeeding. Low adherence to optimal tion, maternal and child health, as well as family plan- ning and other reproductive health issues. The survey breastfeeding including exclusive breast feeding for the first 6months and risk of diarrhoeal disease from con- sample was designed to provide estimates for the entire country, for both urban and rural areas, which com- taminated complementary foods given to infants, well prised of 26 regions from the Tanzania mainland and before 6months of age, is believed to contribute to Zanzibar. under nutrition observed in young children. For instance, the 2010 TDHS reported that 35% of children under-5years of age were stunted while 21% were under Survey design weight indicating that undernutrition is a public health The 2010 TDHS utilised a cross-sectional study design problem in Tanzania that needs to be addressed at a and a nationally representative survey sample was veryearlystage ofinfant’s life.2 obtained using stratified two-stage random sampling.2 In Itis wellestablishedthat optimalbreastfeeding confers the first stage, 475 clusters were selected from a list of protective effects against gastrointestinal infections and enumeration areas from the 2002 Population and improves child survival.3–5 A cohort study carried out in Housing Census.20 Eighteen clusters were selected in Ghana revealed that 22% of neonatal deaths could be each region except Dar es Salaam, where 25 clusters prevented if all infants were put to the breast within the were selected in the Mainland. In Zanzibar, 18 clusters first hour of birth.4 It has also been reported that exclu- were selected in each region for a total of 90 sample sive breastfeeding from birth and until 6months of age points. In the second stage, a complete household hasthepotentialtoprevent13%ofalldeathsamongchil- listing was carried out in each of the selected clusters. dren, aged less than 5years, annually in developing Twenty-two households were selected from each cluster countries.6 in all regions except for Dar es Salaam where 16 house- Research investigating the factors associated with sub- holds were systematically selected. A total of 10300 optimal breastfeeding practices has been conducted in households were selected for the sample, of which developed and developing countries, including African 10176 were successfully interviewed, yielding a house- countries,andshowsthatdelayedinitiationofbreastfeed- hold response rate of 99%. From these households, ingafterbirthandnotexclusivebreastfeedingduringthe 10522 women of reproductive age (15–49years) who first 6months were influenced by factors such as mater- were either permanent residents of the households in nal age,7–9 maternal level of education,8–10 maternal the 2010 TDHS sample or visitors present in the house- employment status,10 11 maternal nutritional status,12 hold on the night before the survey were interviewed. place of delivery,13 14 mode of delivery,15–17 area of resi- Face-to-face interviews were held with the sampled dence,16 household wealth status10 and geographical mothers using a structured questionnaire yielding an 2 VictorR,BainesSK,AghoKE,etal.BMJOpen2013;3:e001529.doi:10.1136/bmjopen-2012-001529 Determinants of breastfeeding indicators in Tanzania interview response rate of 96.4%. Comprehensive details place of residence and geographical zones. The house- regarding the sampling procedure and data collection holdwealth index was calculatedasascoreofhousehold tools areavailable inthe 2010 TDHS report.2 assets weighted using the principal components analysis method.23 Feeding indicators Early initiation of breastfeeding within 1h of birth There are fifteen indicators recommended by WHO19 21 and exclusive breastfeeding were examined in multivari- for assessing infant and young child-feeding practices. ate analysis because their prevalence continues to be The breastfeeding indicators reported in the survey below the national target and WHO/UNICEF recom- include mendation of 90% coverage.24 25 Early initiation of ▸ ‘Early initiation of breastfeeding: proportion of chil- breastfeeding and EBF also playavital role in protecting dren born in the last 24months who were put to the infants against diarrhoeal diseases, and reducing mortal- breast within 1h of birth—this indicator is based on ity among many infants in developing countries.26 The historical recall’. rates of ‘ever-breastfed’ and ‘continued breastfeeding’ ▸ ‘Exclusive breastfeeding under 6months: proportion were very high (>90%); hence, they were not included of infants 0–5months of age who were fed exclusively in multivariate analysis. The indicator for predominant with breast milk—this indicator is based on mother’s breastfeeding was also included in multivariate analysis recall on feeds given to the infant on the previous owing to its impact on increasing the risk of diarrhoeal day’, and respiratory illness in infants.26 In addition, bottle ▸ ‘Continued breastfeeding at 1year: proportion of feeding was not considered in this analysis because the children 12–15months of age who were fed breast prevalencewas very low (4%). milk’. ▸ ‘Continued breastfeeding at 2years: proportion of Data analysis children 20–23months of age who were fed breast Our analysis was restricted to the alive, youngest, last- milk’. born infants aged less than 24months, living with their ▸ ‘Predominant breastfeeding: proportion of infants mothers (women’s age 15–49years) during the 2010 0–5months of age who were fed with breast milk TDHS and the total weighted samplewas 3112. The ana- from the mother (either directly or expressed) and lysis of determinants of early breastfeeding initiation was certain liquids (water, water-based drinks and fruit based on the entire sample (3112 children) whereas juice) and ritual fluids. Infants who received non- those of exclusive breastfeeding and predominant human milk and food-based fluids were not included breastfeeding were based on 837 infants aged from when computing the prevalence ofthis indicator’. 0–5months. Non-exclusive breastfeeding (non-EBF) was ▸ ‘Children ever breastfed: proportion of children born expressed as a dichotomous variable with category 1 for in the last 24months whowereever breastfed’. non-EBF and category 0 for EBF. Delayed initiation of ▸ ‘Bottle feeding: proportion of children 0–23months breastfeeding was expressed as a dichotomous variable breastfeeding of agewho are fedwith a bottle’. with category 0 for early initiation of breastfeeding and WHO recommends that the EBF indicator be disag- category 1 for delayed initiation of breastfeeding. gregated for the following age groups: 0–1, 2–3, 4–5 and Predominant breastfeeding wasexpressedasadichotom- 0–3months. Ever breastfed and early initiation of indica- ous variable with category 1 for predominant breastfeed- tors were further disaggregated and reported for live ing and category 0 for non predominant breastfeeding. births occurring 0–12, 12–23 and 0–23months prior to These variables were examined against a set of interview. It should be noted that the EBF indicator independent variables (individual, household and com- defined above does not represent the percentage of munity characteristics) in order to determine the preva- infants who are exclusively breastfed at 6months of lence and factors associated with delayed initiation of age19 22 but rather the average prevalence of exclusive breastfeeding, non-exclusive breastfeeding and predom- breastfeeding ofchildren <6months ofage. inant breastfeeding indicators. The breastfeeding indicators were examined by the Analyses were performed using Stata V10.0 (Stata individual level factors, which included mother’s Corp, College Station, Texas, USA). ‘Svy’ commands age, mother’s body mass index measured by weight were used to allow for adjustments for cluster sampling (kg)/height (m2), mother’s literacy, mother’s working design, sampling weights and the calculation of standard status, mother’s education, mother’s marital status, part- errors. The Taylor series linearisation method was used ner’s education, partner’s occupation, birth order, birth to estimate CIs around prevalence estimates. A χ2 test interval, sex of child, age of child, perceived size of the was used to test the significance of associations. baby, place of delivery, type of delivery assistance, Unadjusted and adjusted ORs (AOR) were calculated to number of antenatal clinic visits, timing of postnatal estimate the strength of association between independ- check-up, mode of delivery, mother’s access to mass ent variables and three breastfeeding indicator out- media; household level factors included household comes: delayed initiation of breastfeeding, non-exclusive wealth index and mother’s autonomy in household breastfeeding and predominant breastfeeding. In our decision-making, and community level factors included multivariate statistical modelling, wecreated an indicator VictorR,BainesSK,AghoKE,etal.BMJOpen2013;3:e001529.doi:10.1136/bmjopen-2012-001529 3 Determinants of breastfeeding indicators in Tanzania variable for missing data and restricted our analysis to dominated by agricultural activities (63.5%). Of all chil- non-missing data. Multiple logistic regression using dren, more than half (50.9%) were born in health facil- survey commands was conducted using stepwise back- ities but a relatively low percentage of mothers delivered wards elimination of variables in order to determine the by caesarean section (5.1%). Most mothers (48.0%) factors significantly associated with the outcome of were assisted by health professionals at delivery, and a breastfeeding indicators. The ORs with 95% CIs were high proportion of mothers were multiparous (80.4%). calculated in order to assess the adjusted risk of inde- About 48%ofmothers hadmadeatleastthreeantenatal pendent variables, and only those with p<0.05 were clinic visits during pregnancy, and 31.4% had postnatal retained in the final model. We did our backward step- check-ups 41days after birth. The gender of the chil- wise model by adjusting for sampling weights and clus- dren was nearly equally represented in the sample. ters. We double-checked our background elimination Approximately one-quarter (22.6%) of the children method by using the following procedure:(1) enteronly were from poor families. variable with p value <0.20 in our backward elimination process; (2) tested backward elimination by also includ- Breastfeedingindicators ing all variables (all potential confounders) and (3) we Less than half of the mothers (46.1%) had initiated tested for collinearity. The linear interpolation method breastfeeding within the first hour after birth, whereas was used to compute the median duration of exclusive 98.5% reported they had ‘ever breastfed’ their infants breastfeeding. (table 1). 49.9% of infants less than 6months of age Note: in the present study, delayed initiation of breast- were exclusively breastfed, but the median duration for feeding refers to the proportion of children born in the exclusive breastfeeding was only 2.6months. Less than a last 24months who were not put to the breast within 1h quarter (16.8%) of infants below 6months of age were of birth and non-exclusive breastfeeding refers to the predominantly breastfed. About 94% of the children proportion of infants aged 0–5months whowere not fed were still breastfed at 12–15months but the percentage exclusively with breast milk. decreased to 51.1% at 20–23months of age. A few chil- dren (3.8%) werebottle-fed from birth to23months. RESULTS As shown in figure 1, prevalence of EBF was more Basiccharacteristics of the sample than 86% at birth but declined rapidly with age to 23.1% at 6months. At birth, 10.8% of infants were given Online supplementary table S1 shows the distribution of breast milk plus other fluids, including water, juices or 3112 children aged less than 24months according to other milk. individual-, household-level, and community-level char- acteristics. The majority of children lived in rural areas (79.7%). Many mothers (67.0%) had primary level edu- Breastfeedingindicators acrossindividual-level, cation, and about 86% were employed in the last household-level, and community-level characteristics 12months. Eighty-five per cent of the mothers were cur- AsseeninonlinesupplementarytableS3,earlyinitiation rently married, and their husband’s occupation was of breastfeeding within the first hour after birth was Table1 Prevalenceofbreastfeedingindicatorsamongchildrenagedlessthan24monthsinTanzania Sizeofsubsample Indicator (weighted) n(weighted) Rate(%) (95%CI) Earlyinitiationofbreastfeeding0–23months 3112 1434 46.1 (43.44to48.76) Earlyinitiationofbreastfeeding0–11months 1630 750 46.0 (42.58to49.45) Earlyinitiationofbreastfeeding12–23months 1482 685 46.2 (43.00to49.43) Childreneverbreastfed0–23months 3112 3065 98.5 (97.68to99.00) Childreneverbreastfed0–11months 1630 1604 98.4 (97.31to99.07) Childreneverbreastfed12–23months 1482 1461 98.6 (97.63to99.11) Exclusivebreastfeeding0–5months 837 418 49.9 (45.65to54.15) Exclusivebreastfeeding0–1months 245 197 80.7 (74.13to85.94) Exclusivebreastfeeding2–3months 299 153 51.1 (44.38to57.80) Exclusivebreastfeeding4–5months 293 67 22.9 (17.69to29.10) Exclusivebreastfeeding0–3months 544 350 64.4 (58.82to69.66) Predominantbreastfeeding0–5months 837 141 16.8 (13.46to20.74) Continuedbreastfeedingat1year 524 492 94.0 (91.02to96.09) Continuedbreastfeedingat2years 419 214 51.1 (45.19to57.04) Bottlefeeding0–23months 3112 200 3.8 (2.97to4.97) Bottlefeeding0–5months 837 39 4.7 (3.06to7.17) Bottlefeeding6–11months 793 43 5.4 (3.75to7.73) Bottlefeeding12–23months 1482 37 2.5 (1.66to3.83) 4 VictorR,BainesSK,AghoKE,etal.BMJOpen2013;3:e001529.doi:10.1136/bmjopen-2012-001529 Determinants of breastfeeding indicators in Tanzania Figure1 Distributionofchildrenbybreastfeedingstatusaccordingtoage. significantly lower among mothers who delivered at different across individual-level, household-level and home (34%); those who were not assisted by health pro- community-level factors. fessionals (32.7%); residing inruralareas(42.2%); those from poorest households (39.3%) and living in the Determinants of breastfeeding indicators Eastern zone (36.5%) and the Northern zone (39.3%). Unadjusted and adjusted ORs were calculated to estimate There was a significantly lower prevalence of early initi- theeffectofindependentvariablesonthreeinfant-feeding ation of breastfeeding among mothers who delivered by outcomes: delayed initiation of breastfeeding within the caesarean section (20.9%); those who did not have any firsthourafterbirth,non-exclusivebreastfeedingandpre- postnatal check-ups (42.7%); mothers who did not have dominantbreastfeeding.Asseeninonlinesupplementary anyautonomyindecisionmaking(43.1%); motherswho table S4, the adjusted odds of delayed initiation of breast- were unable to read (38.9%) and those with poor access feeding were significantly higher among infants whose to mass media including radio (41.5%) and television mothers were aged less than 24years, had a low level of (44.1%). In contrast, there was a higher prevalence of education (no education/primary education), worked in early initiation of breastfeeding within 1h of birth the last 12months, delivered their babies at home with among mothers from the richest households (62.9%); assistance from an untrained provider (traditional birth from the Central (52.1%), Southern Highland (53.0%) attendants or relatives/other people) and thosewhowere and the Western (51.4%) geographical zones; from delivered by caesarean section. The odds of delayed initi- urban areas (61.2%); who delivered at health facilities ation of breastfeeding were also higher for infants from (57.7%); whoweremarriedto ahusband not involved in rural compared with infants from urban areas. As com- agricultural activities (54.9%), and those who had a pared with infants from the Western geographical zone, higherlevelofeducation(62.7%)andtheirpartnershad infants from Lake, Northern, Eastern and Zanzibar were secondaryandhigherlevelofeducation(62%). at a higher risk of delayed initiation of breastfeeding Exclusive breastfeeding of infants aged less than withinanhourafterbirth. 6months of age was significantly lower among mothers The odds of non-EBF were significantly higher in who had worked in the last 12months (48.0%); mothers infants whose mothers were assisted by traditional birth who resided in urban areas (40.3%); those from the attendants (TBA) at birth than infants of mothers who richest households (37.0%) and those living in Zanzibar were assisted by health professionals. When type of deliv- (10.4%). The proportion of infants who were exclusively ery assistant was removed from the final model and breastfed for the first 6months of life were observed to replaced by place of delivery, we found that place of be higher among mothers from rural areas (52.2%) and delivery was not significantly associated with non-EBF. those living in the Eastern (52.2%), Western (53.6%) Hence, type of delivery assistance was retained in the and Southern (51.6%) geographical zones. The rates of final model. The risk of non-EBF was also significantly predominant breastfeeding were not significantly higher for urban infants compared with their rural VictorR,BainesSK,AghoKE,etal.BMJOpen2013;3:e001529.doi:10.1136/bmjopen-2012-001529 5 Determinants of breastfeeding indicators in Tanzania counterparts. As expected, increasing infant age was highlighting the need to reverse this trend and to associated with significantly low rates of EBF. Infants increase the percentage of initiating breastfeeding from Zanzibar were at greater risk of non-EBF and pre- within 1h of birth. Similarly, the prevalence of EBF in dominant breastfeeding compared with infants from Tanzania was very low in comparison with other neigh- other geographicalzones. bouring African countries such as Uganda (60%),28 Zambia (61%)29 and Malawi (57%).30 A considerable proportion (17%) of infants less than 6months were DISCUSSION predominantly breastfed, suggesting a need for counsel- This study found that less than half of the mothers had ling mothers, caregivers and key family members on the initiated breastfeeding within the first hour after birth, risks associated with predominant feeding. This strategy and that only half of the mothers exclusively breastfed would help to change their behaviour which ultimately their infants aged less than 6months. Seventeen per improved EBF. In our analysis, we found a significant centof infants lessthan 6months ofagewerepredomin- association between maternal young age (15–24years) ately breastfed. We found that lower maternal education, and delayed initiation of breastfeeding. This result is younger maternal age, being employed, delivered at consistent with findings from India which showed that home, delivered by caesarian section, delivery assistance older mothers (≥35years) were at lower risk of delayed by untrained provider and residing in rural areas of initiation of breastfeeding compared with young Eastern, Lake, Northern and Zanzibar were determi- mothers (AOR for older mothers ≥ 35years= 0.72, 95% nants of delays in initiation of breastfeeding within first CI 0.50 to 1.02).13 We further explored this association hour after birth. Similarly, delivery assistance by an with parity and found that most of theyoung mothers in untrained provider and residing in urban areas of Tanzania were first-time mothers, suggesting that they Zanzibar were predictors of non-exclusive breastfeeding lacked knowledge or experience about appropriate in the first 6months of infant’s life. We have also identi- breastfeeding practices. Hence, the need for health pro- fied the target groups of women who need more breast- fessionals and traditional birth attendants to provide feeding support that included young, uneducated, adequate support to encourage young and first-time employed women <25years, women from both rural and mothers to establish early initiation of breastfeeding urban areas in the Eastern, Lake, Northern and within 1h after giving birth. Zanzibar geographical zones and women who also Similar to the findings reported in India,13 we also lacked propercareduring and after birth. found that women with higher levels of education had a This paper is one of the few reports from Africa, reduced risk of delayed initiation of breastfeeding and including Tanzania, which has described the prevalence this might be explained by their exposure to various of breastfeeding practices using the most recent nation- sources of information and better knowledge about ally representative data from Tanzania, and the current appropriate infant and young child feeding. The varia- WHO-recommended definitions for assessing Infant and tions in the prevalence of early initiation of breastfeed- Young Child Feeding indicators. The findings from this ing across different geographical zones could be due to study will help guide health programmes to improve cultural differences and taboos about breastfeeding new- early initiation of breastfeeding, and exclusive breast- borns with first breast milk (ie, colostrum) in different feeding in order to ensure young children in Tanzania regions of Tanzania.31–33 A majorconcern isthe verylow receive the full benefits of appropriate breastfeeding prevalence of EBF in Lake and Zanzibar geographical practices including reduced morbidityand mortality. zones. This could be due to inadequate knowledge The main strengths of this study include the use of a among mothers and family members regarding benefits large nationally representative survey sample, with very of exclusive breastfeeding in the first 6months of high response rate to the survey interviews (96.4%), infant’s life and also existence of belief that breast milk comprehensive data on standard infant-feeding indica- alone is not sufficient to fulfil infant’s hunger hence tors to identify factors associated with suboptimal breast- complement with other liquids/soft foods.31 Provision of feeding practices in Tanzania, and appropriate sampling adequate support and educating mothers and their fam- design in the analysis. Furthermore, we restricted the ilies from these zones on the importance of giving initial sample for children to only those who lived with their breast milk to infants and EBF until 6months may have mothers to ensure greater accuracy of information a positive effect on improving rates of early initiation of regarding breastfeeding practices. The main limitation breastfeeding and EBF and potentially reduce the risks was the cross-sectional nature of the survey which of infections and death among newborns.4 limited inferences about causality from the analyses. In In this study, rural infants had significantly higher risk addition, EBF was based on a 24h recall rather than a of delayed initiation of breastfeeding within 1h after longer recall period, and this short recall may have birth compared with urban infants. This finding is in missed some infants who were fed other liquids or foods agreement with the previous studies from the Morogoro prior to24hbeforethe survey. region in Tanzania31 and from Ethiopia.34 The differ- The prevalence of early initiation of breastfeeding has ence in early initiation of breast feeding between rural declined from 59% in 200527 to 46% in 2010, and urban mothers might be explained by the high 6 VictorR,BainesSK,AghoKE,etal.BMJOpen2013;3:e001529.doi:10.1136/bmjopen-2012-001529 Determinants of breastfeeding indicators in Tanzania percentage of rural women who delivered at home examine the relationship between maternal working (93%) assisted by TBAs and other people such as family status and area of residence found that, most working members. These birth attendants may have had inad- mothers (68.6%) resided in rural areas than in urban equate knowledge of the benefits of this feeding practise areas (12.5%), and they had higher risk of delayed initi- and thus failed to support mothers in initiating breast- ation of breastfeeding within 1h after birth, as discussed feeding early. Furthermore, negative cultural beliefs earlier. We also found that the prevalence of EBF about colostrum and lower level of education among decreased with increasing age of the child. This finding rural mothers (88%) might also have contributed. Rural was in conformity with other secondary analyses of women may need more support to overcome barriers to demographic and health surveys conducted in Nigeria,18 early initiation of especially those living in the Eastern, India,13 Bangladesh,39 Sri Lanka, Cambodia, Indonesia, Lake, Northern and Zanzibar geographical zones. On Philippines and Timor-Leste, and Vietnam9 and the other hand, mothers from urban areas were at Malawi14 which have also reported a declining preva- greater risk of poor EBF practices than mothers from lence ofEBFasthe age of the child increased. rural areas, possibly because of the demand to return to work after maternity leave11 since most of these urban mothers were in paid employment. Also, most mothers in urban areas were from families with higher socio- CONCLUSIONS economic status compared with rural areas and that may The prevalence of breastfeeding indicators regarding have facilitated access to breast milk substitutes. For early initiation of breastfeeding and EBF were below the example, data from a multilevel analysis of factors asso- nationaltargets (90% coverage)40 and improvement is ciated with non-EBF in nine East and Southeast Asian needed, for infants, to gain the full benefits of breast- countries revealed that improved socioeconomic status feeding and help the country achieve the Millennium both at individual levels and community levels was a Development Goal for reduction of infant mortality negativefactor for EBF.9 from 51 deaths per 1000 births in 2010 to 38 deaths per The risk of delayed initiation of breastfeeding in the 1000 live births by the year 2015.2 The improvement of first hour after birth was significantly higher among breastfeeding practices will require national level pro- mothers who delivered at home compared with those grammes with a focus on target groups with suboptimal who delivered at health facilities. Similarly,having a baby breastfeeding practices including young, uneducated not delivered by a health professional was a significant mothers who deliver at home assisted by untrained predictorof non-EBF. This indicatesthe need to educate health personnel, and those who deliver by caesarean key family members and TBAs about the benefits of section. Further research is recommended to investigate initial breast milk for the newborn so that they can why early initiation of breastfeeding is decreasing in encourage mothers who deliver at home to establish Tanzania. breast-feeding immediately after birth and EBF up to 6months. Exclusive breastfeeding should also be pro- Authoraffiliations moted at health facilities during antenatal care visits and 1SchoolofHealthSciences,UniversityofNewcastle,Newcastle,NewSouth during deliveries; and at the community level through Wales,Australia peercounselling support for EBF.35 2SchoolofMedicine,UniversityofWesternSydney,Newcastle,NewSouth Wales,Australia Delivery by caesarean section was a risk factor for 3SydneySchoolofPublicHealth,SydneyMedicalSchool,TheUniversityof delayed initiation of breastfeeding in Tanzania. This Sydney,Newcastle,NewSouthWales,Australia finding is consistent with previous reports from India,13 Nepal36 and Sri Lanka.37 This association may be linked Acknowledgements WeacknowledgetheAustralianAgencyforInternational to the effects of anaesthesia delaying the onset of lacta- DevelopmentforsponsoringtheMPhilfellowshipforRV. tion and some baby-unfriendly postoperative-care prac- Contributors RVdesignedthestudy,performedtheanalysisandpreparedthe tices.38 A recent systematic review and meta-analysis of manuscript;SKBprovidedadviceonstudydesign,andcriticallyrevisedthe observational studies that examined influence of caesar- manuscriptforintellectualcontent;KEAprovidedadviceonstudydesign, data-analysisandcriticallyrevisedthemanuscriptforintellectualcontent;and ean delivered on early breastfeeding showed that caesar- MJDprovidedadviceonstudydesignandcriticallyrevisedthemanuscriptfor ean delivery has a significant adverse association with intellectualcontent. early breastfeeding.17 Appropriate guidelines for caesar- Funding Thisresearchreceivednospecificgrantfromanyfundingagencyin ean deliveries are needed to minimise delays in initi- thepublic,commercialornot-for-profitsectors. ation of breastfeeding. Prospective mothers and health Competinginterests None. workers should be informed about the negative associ- ation between prelabour caesarean delivery and breast- Ethicsapproval TheethicsapprovalwasobtainedfromtheUniversityof feeding and the implications for infant well-being.17 NewcastleHumanResearchEthicsCommittee. Our analysis showed a negative association between Provenanceandpeerreview Notcommissioned;externallypeerreviewed. maternal working status and early initiation of breast- Datasharingstatement Appendicestotheextendedreportareavailablein feeding 1h after birth. However, our subanalysis to English. VictorR,BainesSK,AghoKE,etal.BMJOpen2013;3:e001529.doi:10.1136/bmjopen-2012-001529 7 Determinants of breastfeeding indicators in Tanzania REFERENCES 21. WorldHealthOrganization.Indicatorsforassessinginfantandyoung 1. WorldHealthOrganization.Globalstrategyforinfantandyoungchild childfeedingpractices.Part3CountryProfile.Geneva:WHOPress, feeding.Geneva:WorldHealthOrganization,2003. 2010. 2. NationalBureauofStatistics(NBS)(Tanzania)andICFMacro. 22. AgampodiSB,AgampodiTC,SilvaA.ExclusivebreastfeedinginSri TanzaniaDemographicandHealthSurvey2010.NBSandICF Lanka:problemsofinterpretationofreportedrates.IntBreastfeedJ Macro.DaresSalaam,Tanzania,2011. 2009;4:14. 3. WHOcollaborativestudyteamontheroleofbreastfeedingonthe 23. GwatkinDR,RutsteinS,JohnsonK,etal.Socioeconomic reventionofinfantmortality.Effectofbreastfeedingoninfantand differencesinhealth,nutritionandpoverty.Washington,DC:HNP/ childmortalityduetoinfectiousdiseasesinlessdeveloped PovertyThematicGroupoftheWorldBank,2000. countries.Apooledanalysis.Lancet2000;355:451–5. 24. MinistryofHealthandSocialWelfareandTanzaniaFoodand 4. EdmondK,ZandohC,QuigleyMA,etal.Delaybreastfeeding NutritionCentre.Tanzanianationalstrategyoninfantandyoung initiationincreasesriskofneonatalmortality.Pediatrics2006;117: childnutrition.MinistryofHealth,DaresSalaam,Tanzania,2004. e380–6. 25. WHO/UNICEF.Globalactionplanforpreventionandcontrolof 5. EdmondKM,KirkwoodBR,Amega-EtegoS,etal.Effectofearly pneumonia(GAPP).2009.http://www.unicef.org/media/files/ infantfeedingpracticesoninfection-specificneonatalmortality:ana GAPP3_web.pdf(accessed14March2012). investigationofcausallinkswithobservationaldatafromGhana.Am 26. BlackRE,AllenLH,BhuttaZA,etal.Martenalandchild JClinNutr2007;86:1126–31. undernutrition:globalandregionalexposuresandhealth 6. JonesG,SteketeeRW,BlackRE,etal.Howmanychilddeathscan consequences.Lancent2008;371:243–60. wepreventthisyear?Lancet2003;362:65–71. 27. NationalBureauofStatistics(NBS)TanzaniaandORCMacro. 7. MelinaT,MariaS,AntonisG,etal.BreastfeedinginAthens, TanzaniaDemographicandHealthSurvey2004–05.NationalBureau Greece:factorsassociatedwithitsinitiationandduration.JPed ofStatisticsandORCMacro.DaresSalaam,Tanzania,2005. GastrNutr2006;43:379–84. 28. UgandaBureauofStatistics(UBOS)MacroInternationalInc. 8. LandeB,AndersenLF,BaerugA,etal.Infantfeedingpracticesand UgandaDemographicandHealthSurvey2006.Calverton,MD: associatedfactorsinthefirstsixmonthsoflife:theNorwegianinfant UBOSandMacroInternationalInc,2007. nutritionsurvey.ActaPaediatr2003;92:152–61. 29. CentralStatisticalOffice(CSO),MinistryofHealth(MOH),Macro 9. SenarathU,DibleyMJ,AghoKE.Factorsassociatedwith International.ZambiaDemographicandHealthSurvey2007. nonexclusivebreastfeedingin5EastandSoutheastAsiancountries: Calverton,MD:CSOandMacroInternationalInc,2009. amultilevelanalysis.JHumLact2010;26:248. 30. NationalStatisticalOffice(NSO)andORCMacro.Malawi 10. AmirL,DonathS.Socioeconomicstatusandratesofbreastfeeding DemographicandHealthSurvey2004.Calverton,MD:NSOand inAustralia:evidencefromthreerecentnationalhealthsurveys.Med ORCMacro,2005. JAust2008;189:254–6. 31. ShirimaR,GreinerT,KylbergE,etal.Exclusivebreast-feedingis 11. LakatiA,BinnsC,StevensonM.Theeffectofworkstatuson rarelypractisedinruralandurbanMorogoro,Tanzania.Public exclusivebreastfeedinginNairobi.AsiaPacJPublicHealth HealthNutr2001;4:147–54. 2001;14:85–90. 32. SellenDW.InfantandyoungchildfeeingpracticesamongAfrican 12. LiuJ,SmithMG,DobreMA,etal.Maternalobesityand pastoralist.JBiosocSci1998;30:481–99. breastfeedingpracticesamongwhiteandblackwomen.Obesity 33. AgnassonI,MpelloA,GunnLaugssonG,etal.Infantfeeding 2009;18:175–82. practicesduringthefirstsixmothsoflifeinaruralareaofTanzania. 13. PatelA,BadhoniyaN,KhadseS,etal.Infantandyoungchild EastAfrMedJ2001;4:9–13. feedingindicatorsanddeterminantsofpoorfeedingpracticesin 34. SetegnT,GerbabaM,BelachewT.Determinantsoftimelyinitiation India:secondarydataanalysisofNationalFamilyHealthSurvey ofbreastfeedingamongmothersinGobaWoreda,SouthEast 2005–06.FoodNutrBull2010;31:314–33. Ethiopia:across-sectionalstudy.BMCPublicHealth2011;11:217. 14. VaahteraM,KulmalaT,HietanenA,etal.Breastfeedingand 35. TylleskarT,JacksonD,MedaN,etal.Exclusivebreastfeeding complementaryfeedingpracticesinruralMalawi.ActaPaediatr promotionbypeercounselorsinsub-SaharanAfrica(PROMISE-EBF): 2001;90:328–32. aclusterrandomisedtrial.Lancet2011;378:420–7. 15. HauckYL,FenwickJ,DhaliwalSS,etal.AWesternAustralian 36. PandeyS,TiwariK,SenarathU,etal.Determinantsofinfantand surveyofbreastfeedinginitiation,prevalenceandearlycessation youngchildfeedingpracticesinNepal:secondarydataanalysisof patterns.MaternChildHealthJ2011;15:260–8. DemographicandHealthSurvey2006.FoodNutrBull 16. DibleyMJ,RoySK,SenarathU,etal.Across-countrycomparisons 2010;31:334–51. ofselectedinfantandyoungchildfeedingindicatorsandassociated 37. SenarathU,DibleyMJ,GodakandageSSP,etal.Determinantsof factorsinfourSouthAsiancountries.FoodNutrBull infantandyoungchildfeedingpracticesinSriLanka:secondary 2010;31:366–74. dataanalysisofDemographicandHealthSurvey2000.FoodNutr 17. PriorE,SanthakumaranS,GaleC,etal.Breastfeedingafter Bull2010;31:352–65. caesareandelivery:asystematicreviewandmeta-analysisofworld 38. MathewsMK.Therelationshipbetweenmaternallabouranalgesia literature.AmJClinNutr2012;95:1113–35. anddelayininitiationofbreastfeedinginhealthyneonatesinthe 18. AghoKE,DibleyMJ,OdiaseJI,etal.Determinantsofexclusive earlyneonatalperiod.Midwifery1989;5:3–10. breastfeedinginNigeria.BMCPregnancyChildbirth2011;11:2–8. 39. MihrshahiS,KabirI,RoySK,etal.Determinantsofinfantand 19. WorldHealthOrganization.Indicatorsforassessinginfantandyoung youngchildfeedingpracticesinBangladesh:secondarydata childfeedingpractices.PartI:Definitions.Conclusionofa analysisofDemographicHealthSurvey2004.FoodNutrBull consensusmeetingheld6–8November2007inWashington,DC, 2010;31:295–313. USA,2008. 40. MinistryofHealthTanzaniaandTanzaniaFoodandNutrition 20. NationalBureauofStatistics(NBS)Tanzania.2002-populationand Centre.Tanzanianationalstrategyoninfantaandyoungchild housingcensus.DaresSalaam;2002,GeneralReport. nutritionimplementationplan.MinistryofHealth,DaresSalaam, Tanzania,2004:6 8 VictorR,BainesSK,AghoKE,etal.BMJOpen2013;3:e001529.doi:10.1136/bmjopen-2012-001529

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.