Determinants of exclusive breastfeeding in Tanzania: a systematic review and meta-analysis Authors: Mariam Noorani, Rukhsar Osman, Muzdalifat Abeid, Jai Das, Zulfiqar Bhutta Published: 2025 ABSTRACT AbstractBackgroundExclusive breastfeeding (EBF) is a critical intervention for reducing childhood morbidity and mortality, yet global EBF rates remain below targets. In Tanzania, EBF rates vary widely across regions, with reported disparities influenced by individual, familial, and health system factors. Despite increasing national prevalence rates of EBF from 41% in 2005 to 64% in 2022, localized studies show lower prevalence ranging from 20 to 30%, highlighting the need for a comprehensive synthesis of evidence on the determinants of EBF to inform policy in Tanzania.MethodsWe searched PubMed, Scopus, and CINAHL for peer-reviewed studies from 2014 to 2024 assessing EBF determinants among Tanzanian infants aged less than 6 months. Data was critically appraised using Joanna Briggs Institute tools, and synthesized via random-effects meta-analysis to pool the Odds Ratios of determinants of EBF. The review is registered with PROSPERO: CRD42024566722ResultsA total of 183 studies were identified and 130 underwent title and abstract screening. Fourteen were retrieved for full text screening and 11 met inclusion criteria and were included for analysis. Four studies were among HIV positive mothers and six studies were from Northern Tanzania. Meta analysis of 14 determinants revealed that married/cohabiting mothers had significantly higher odds of EBF (OR 1.63, 95% CI: 1.12,2.36) compared to single mothers. Sub-group analysis among HIV positive mothers showed that employed mothers had higher odds of EBF (OR 1.37, 95% CI: 1.10,1.72).ConclusionMarital status is a key determinant of EBF in Tanzania, underscoring the role of interpersonal and household support. To accelerate progress toward the 70% EBF target, targeted lactation support should be given to the vulnerable groups including single mothers. At national level, interventions should address regional disparities, engage families and communities in breastfeeding education and strengthen infant feeding counselling for HIV positive mothers. Future research should expand coverage to understudied regions and employ qualitative approaches and longitudinal designs to strengthen the evidence base.Supplementary informationThe online version contains supplementary material available at 10.1186/s13006-025-00790-w.Keywords: Exclusive breastfeeding, Determinants, Tanzania, Meta-analysis, Systematic review, Factors, Infants Background Breastfeeding is recognized as an important intervention to promote child health, development and survival because of its impact in reducing childhood morbidity and mortality. The health benefits of breastfeeding range from reduced risk of infections as well as chronic illnesses like autoimmune diseases, allergic conditions, cardiovascular diseases and leukemia [1]. Globally, an estimated 820,000 deaths of children under the age of five could be avoided every year if breastfeeding is practiced as per the recommended international guidelines [2]. Despite well-known advantages, exclusive breastfeeding (EBF) rates in the first six months of life are only 48% globally [3]; well below the adjusted target of 70% by the year 2030 [4] There exists a significant variation in breastfeeding rates across various continents and countries. In Africa for example, exclusive breastfeeding (EBF) rates in infants aged 0–5 months range from 80% in Rwanda to less than 20% in Gabon [5]. In East Africa, both Rwanda and Burundi have higher rates; over 80% compared to Kenya, Uganda and Tanzania; whose EBF rates are between 60 and 70%. Within a single country, the breastfeeding practices also vary in different regions; in Tanzania, highest rates are found in the Lake region and lowest in the Zanzibar islands [6]. Determinants of EBF practice can range from individual characteristics to family and community related variables to policy level determinants. In the African context maternal age, education level, wealth quintile, antenatal care visits and early initiation of breastfeeding have been found to influence EBF rates [7]. Systematic reviews from different African countries have identified contextual determinants that can guide national efforts and policies [8–10]. Tanzania has seen a steady increase in rates of EBF from 41% in 2005 to 64% in 2022 as reported by the Tanzania Demographic and Health Survey (TDHS) [6] and has been in the frontline to promote breastfeeding practices. However, individual studies show much lower prevalence of 20–30% in some parts of the country [11–13]. A few studies have investigated the determinants of exclusive breastfeeding [13–21], but a preliminary search of the literature revealed that no systematic reviews or meta-analyses have been conducted to synthesize this data. Therefore, this study aimed to conduct a systematic review and meta-analysis of the determinants of exclusive breastfeeding among infants younger than six months in Tanzania. A review and synthesis of the evidence on determinants of the practice of EBF will guide policy and influence the design of pragmatic interventions in the country. Methods This systematic review and meta-analysis was guided by the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) recommendations [22]; (checklist attached as Additional file 1) with the registration of the protocol with PROSPERO International Prospective Register of Systematic Reviews (registration no: CRD42024566722). The full review protocol is also available at this site. Eligibility criteria We included all peer-reviewed primary studies conducted in Tanzania that assessed the predictors, determinants or factors associated with EBF that were published over a 10-year period (January 2014 to July 2024). This time period was selected to capture more recent studies post Millennium Development Goals (MDGs) era and focusing on the Sustainable Development Goals (SDG) period. Qualitative studies, studies reporting secondary data analysis and those that assessed only prevalence of EBF without determinants were excluded. The study population was children aged 0 to 6 months whose feeding information was available and the outcome of interest was exclusive breastfeeding which was defined as being fed only breast milk without any other liquids or solids with the exception of prescribed medications. Classification of determinants was based on the Social-ecological model of health into individual, interpersonal, community and health services variables [23]. This approach has been described previously to prioritise interventions targeting determinants at different levels of the mother infant ecosystem. The Lancet 2023 breastfeeding series described how breastfeeding practices are influenced by individual mother-infant attributes which in turn are determined by the settings in which they live and the influences from external structures including policies [24] In our systematic review; individual maternal and infant factors included: age of mother and infant, level of education of mother, parity, knowledge and perceptions of EBF, employment status, mode of delivery, birth weight and breastfeeding problems. Interpersonal factors were marital status, age and education level of spouse/partner, type of family, income and residence (rural or urban). Health service related factors included place of delivery, type of health facility, breastfeeding counseling, initiation of breastfeeding and number of antenatal visits. Community level factors were socio-economic status, workplace challenges, breastfeeding education and support during antenatal and postnatal care. Search strategy We searched PubMed, Scopus and CINAHL using a combination of subject headings and free text terms. PubMed search was completed on 26th July 2024, Scopus on 6th August 2024 and CINAHL on 17th August 2024. The database searches were carried out using the following terms for population of interest: “Infant”, “Newborn”, “Baby”; combined with outcome terms: “Breast Feeding”, “exclusive breastfeeding”, “Breast Milk Expression”, “Lactation”, “pumped milk”, “expressed milk”; while terms for exposure of interest were: “determinants”, “factors”, “barrier”, “predictor”, “associated factor”, “challenge”, “facilitator”. These were combined with “Tanzania”. The search was first developed for the PubMed database then translated for the other databases using subject headings suitable for each database’s thesaurus. Boolean and proximity operators were used to combine synonyms and keywords describing the main concepts, which were then searched as free-text terms. To include variations of the terms, truncations and wildcards were used. We finally conducted a manual search for additional studies in relevant bibliographies of key papers. Selection of published studies and data extraction Articles that were identified from the search of different databases were exported to the Covidence software [25] for screening and further review. The titles and abstracts were screened independently by the first 2 authors and full texts of the relevant articles were obtained. Both authors independently read the full texts and those that met the pre-defined eligibility criteria were retained for quality appraisal and data extraction. RO extracted information on the study title, authors, year of publication, study duration, study location, design, sample size, sampling technique, study participants, their inclusion and exclusion, measurement of the outcome variable of EBF and factors associated with EBF. The data was independently reviewed and cross-checked with the full-texts by MN for accuracy and completeness. Any disagreements during the review process were resolved through discussions and consensus among the reviewers. Critical appraisal The quality of methodology of the studies reviewed was evaluated using the Joanna Briggs Institute’s (JBI) critical appraisal tools [26]. The JBI appraisal tools are intended to comprehensively evaluate the methodological quality of studies and ascertain the extent to which each study addressed the potential for bias in its design, conduct, and analysis. A variety of JBI appraisal tools are available for the quality assessment of different studies; we used appraisal tools for analytical cross-sectional studies and cohort studies. The first 2 authors independently did the quality assessment and consensus was used to resolve any disagreements. The cross-sectional appraisal tool has eight items, while the tools for cohort studies consist of 11 items. In this study, the original tools were applied without modification. “Yes” or “No” were used to indicate whether a criterion was present or absent, respectively, while “unclear” was used when authors did not provide sufficient details to allow for criterion evaluation. The appraisal tool does not have distinct cut-off points, therefore the authors used cut-off values from previously reported literature [9]; for cross-sectional studies, we considered scores of eight or seven, six to four, and three or less as low, moderate or high risk of bias, respectively. Scores of eleven or ten, five to nine, and four or less were considered low, moderate, and high risk of bias in cohort studies, respectively. Data synthesis and analysis Regrouping of categories of variables was done to allow for comparable analysis across studies. The new groups were determined based on the available data as well as acceptable categorizations from other datasets including the Demographic and health surveys (DHS). Maternal age was grouped into < 35 years and ≥ 35 years. Education was categorized as primary level or less and secondary level or higher. Parity and gravidity was either Primipara/primigravida or multipara/multigravida. Marital status was described in different studies as single, married, cohabiting, widowed, divorced. This was regrouped into married/cohabiting and single. Employment status was described in some studies as “yes” and “no” while others categorized as “type of employment”. This was regrouped into employed or not employed. Number of antenatal visits were re-categorized into adequate if 4 or more visits were done and inadequate if less than 4 visits. To obtain the pooled Odds Ratios (OR) and 95% CI, extracted data on the various determinants of EBF were fitted into a random effects model using inverse variance. For the meta-analysis, we included factors which had ORs reported in at least 2 studies and presented these as forest plots. We conducted a subgroup analysis of determinants of EBF in the studies that had a population of HIV positive mothers and their infants. The level of significance was set at p < 0.05 and the meta-analysis was done on RevMan software [27]. The I2 statistic was calculated for each determinant to assess heterogeneity. Results The search of electronic databases identified 183 studies. After removal of duplicates (n = 53) and screening of titles and abstracts, 14 full texts were eligible for review. Of these, 4 were excluded; 3 of them did not report EBF as the outcome while 1 of them was a multi-country study which did not include data from individual sites. A manual search of the citations and references of the 10 eligible studies identified 1 more relevant study; hence 11 studies were included in the review (Fig. 1). Prisma flow diagram showing search process and study selection Characteristics of included studies Seven of the studies used a cross-sectional design and the remaining four were prospective cohort studies. The sample sizes ranged between 128 and 1128 participants. The population of four of the studies was HIV positive mothers [16–18, 21]. There were five studies conducted at health facilities and at community settings each, while 1 study had participants from both settings. Six studies were conducted in the Northern region of Tanzania, three were from the Coastal regions, and one study each was from the Southern highlands and Zanzibar. EBF was recorded in different ways: by recall since birth in 7 studies, one study documented age at introduction of non breast milk foods, one study used both recall since birth and 24 hour recall, 1 study used both 3 day recall and prospective daily log and in 1 study it was not clear how the outcome was measured [19]. Eight studies reported both crude and adjusted ORs while three studies had only crude OR. Table 1 summarises characteristics of the studies included. Table 1Characteristics of studiesAuthor, yearStudy location and settingStudy design, sample size and sampling techniqueStudy participantsMeasurement of EBF variableDeterminants of EBFMethod of analysisQuality rating (risk of bias)Faustine, 2022• Southern highlands of Tanzania (Njombe and Iringa regions)• Health facility• Cross-sectional study• 372 participants• Multistage sampling• (census method, lottery method, systematic random sampling)HIV-positive lactating mothers with an infant aged 6 to 12 months attending PMTCT program,Recall from birth to 6 months• Knowledge about EBF• Adequate antenatal visits• Positive perception about EBF• Income above 1$ per day• Never having experienced breast problemCrude OR and Adjusted OR(Adjusted for level of education)ModerateHanselman, 2018• Northern Tanzania (Manyara region)• Community• Cohort study• 250 participants• Sampling technique not clearMothers with healthy singleton newborn; weight greater than 1500 gramsAge at introduction of non-breastmilk foods• Higher sanitation score increased the likelihood of non-breastmilk feed’s introduction before 60 days• Farmers owning cattle were more likely to introduce animal milk before 60 days.Crude ORModerateTamara, 2019• Northern Tanzania (Kilimanjaro region)• Health facility• Cohort study• 364 participants• Purposive samplingMother-infant pairs from delivery up to six months.24 hour recall and recall since birthWe have reported 24 hour recall as per TDHS• Unemployed mothers,• Younger mothers (less than 35 yrs)Crude OR and Adjusted OR (adjusted for education, marital status, parity, antenatal visits, breastfeeding knowledge)LowKazaura, 2016• Eastern zone (Coastal region)• Community• Cross-sectional study• 342 participants• Random samplingMothers of children aged 6 months to 23 monthsRecall since birth• Mothers with education level above primary• Positive attitude towards breastfeedingCrude and adjusted Ors (adjusted for maternal occupation)ModerateKhamis, 2017• Zanzibar• Community• Cross sectional study• 303 participants• Simple random samplingMother-infant pairs less than 6 monthsCurrent feeding practices• Young mothers (20-25 yr)• Hospital delivery• Strong community/madrasa supportCrude and Adjusted ORs (Adjusted for number of children, time mother is away from child, knowledge of BF)HighMadsen, 2019• Northern Tanzania• (Kilimanjaro region)• Health facility and community• Cohort study• 1128 participants• Census samplingWomen with live singleton birthRecall since birthExposure to Intimate partner violence is associated with increased risk of premature termination of EBF.Stratified analysisLowMaonga, 2016• Northeastern Tanzania (Tanga region)• Community• Cross- sectional study• 316 participants• Multistage samplingWomen with infants aged 6–12 months,breastfeeding in the first days of life.Recall since birth• Older mothers (≥35 years)• Good knowledge on duration and advantages of EBFAdjusted OR (adjusted for employment, Advice during ANC on breastfeeding, advice on breastfeeding during postnatal period)ModerateMoshi, 2021• Northern Tanzania (Manyara region)• Community• Cross-sectional study• 342 participants• Multistage technique and simple random samplingMothers with infants 0–6 months of ageNot clear• Older maternal age• Infants of younger age (0-1 month)• Mothers with at least primary level of education• Adequate knowledge on EBFCrude and adjusted Ors (adjusted for maternal occupation, number of ANC visits and place of delivery)ModeratePhilemon. 2022• Northern Tanzania• (Kilimanjaro region)• Health facility• Cross sectional study• 524 participants• Sampling technique not clearMothers with children less than 2 years of age who were enrolled in selected PMTCT clinics.Recall since birth• Mothers who had had more than one pregnancy while HIV-positiveNegative associations:• Mothers diagnosed with HIV for more than 5 years• Having ≥4 ANC visits• ANC from dispensary or health centres as compared to hospitalsCrude and adjusted ORsModerateRasheed, 2018• Northern Tanzania• (Kilimanjaro region)• Health facility• Cross sectional study• 128 participants• Census samplingMothers with HIV-exposed infants aged 6 to 12 monthsRecall since birthNo statistically significant findingsCrude ORsModerateWilliams, 2016• Eastern zone• (Coastal region)• Health facility• Cohort study• 226 participants• Sampling technique not clearHIV infected mothers with children less than 18 months3-day recall and prospective daily log• Child age• Enrolment facility• Disclosure to maternal grandmother• Lower level of maternal educationCrude and adjusted Ors (adjusted for maternal age, Maternal HIV treatment, infant HIV status, marital status, food insecurity, asset score)Moderate Characteristics of studies • Southern highlands of Tanzania (Njombe and Iringa regions) • Health facility • Cross-sectional study • 372 participants • Multistage sampling • (census method, lottery method, systematic random sampling) • Knowledge about EBF • Adequate antenatal visits • Positive perception about EBF • Income above 1$ per day • Never having experienced breast problem Crude OR and Adjusted OR (Adjusted for level of education) • Northern Tanzania (Manyara region) • Community • Cohort study • 250 participants • Sampling technique not clear • Higher sanitation score increased the likelihood of non-breastmilk feed’s introduction before 60 days • Farmers owning cattle were more likely to introduce animal milk before 60 days. • Northern Tanzania (Kilimanjaro region) • Health facility • Cohort study • 364 participants • Purposive sampling 24 hour recall and recall since birth We have reported 24 hour recall as per TDHS • Unemployed mothers, • Younger mothers (less than 35 yrs) • Eastern zone (Coastal region) • Community • Cross-sectional study • 342 participants • Random sampling • Mothers with education level above primary • Positive attitude towards breastfeeding • Zanzibar • Community • Cross sectional study • 303 participants • Simple random sampling • Young mothers (20-25 yr) • Hospital delivery • Strong community/madrasa support • Northern Tanzania • (Kilimanjaro region) • Health facility and community • Cohort study • 1128 participants • Census sampling • Northeastern Tanzania (Tanga region) • Community • Cross- sectional study • 316 participants • Multistage sampling Women with infants aged 6–12 months, breastfeeding in the first days of life. • Older mothers (≥35 years) • Good knowledge on duration and advantages of EBF • Northern Tanzania (Manyara region) • Community • Cross-sectional study • 342 participants • Multistage technique and simple random sampling • Older maternal age • Infants of younger age (0-1 month) • Mothers with at least primary level of education • Adequate knowledge on EBF • Northern Tanzania • (Kilimanjaro region) • Health facility • Cross sectional study • 524 participants • Sampling technique not clear • Mothers who had had more than one pregnancy while HIV-positive Negative associations: • Mothers diagnosed with HIV for more than 5 years • Having ≥4 ANC visits • ANC from dispensary or health centres as compared to hospitals • Northern Tanzania • (Kilimanjaro region) • Health facility • Cross sectional study • 128 participants • Census sampling • Eastern zone • (Coastal region) • Health facility • Cohort study • 226 participants • Sampling technique not clear • Child age • Enrolment facility • Disclosure to maternal grandmother • Lower level of maternal education Critical appraisal of studies Of the 7 cross- sectional studies, 6 of them had a moderate risk of bias and only 1 had a high risk of bias. Most of them did not address confounders effectively. Of the 4 cohort studies, 3 had a moderate risk of bias while one had a low risk of bias. The risk of bias for each study is provided in Table 1 and the individual study ratings are available as Additional file 2. Meta-analysis of factors associated with EBF Meta-analysis was done for 14 determinants of EBF which were reported in at least 2 studies. As shown in the forest plot in Fig. 2, the only determinant that showed statistical significance was the marital status of the mother with mothers who are married or cohabiting having higher odds of EBF as compared to those who are single (OR 1.63, 95% CI: 1.12, 2.36, p-value: 0.03). Forest plot showing pooled or for marital status of the mother There were no significant findings for any of the other determinants of EBF; however, maternal education had a pooled OR of 0.61 (95% CI:0.36,1.02) suggesting that mothers with primary education or less had lower odds of EBF compared to secondary or higher. Maternal employment status (Pooled OR: 1.24, 95% CI: 0.57, 2.70), number of antenatal visits (Pooled OR: 0.72, 95% CI: 0.35, 1.47) and maternal age (Pooled OR: 0.99, 95% CI: 0.35, 2.79) did not show statistical significance in the meta-analysis despite demonstrating significance in individual studies. A summary of the pooled ORs for the other determinants is shown in Table 2 and individual forest plots are provided as Additional file 3. As shown in the table, from the calculated I2 statistic, there was significant heterogeneity observed in many of the effect sizes; reflecting not only possible methodological differences in the studies but also true differences due to different populations and samples being studied. Table 2Summary of pooled ORs for determinants of EBFVariableNumber of studiesCategoriesOdds ratio (95% CI)HeterogeneityI2P valueMaternal age4 < 35 years≥35 years0.99 (0.35, 2.79)75%0.009Parity3PrimiparaMultipara0.90 (0.47, 1.70)0%0.44Gravidity2PrimigravidaMultigravida0.89 (0.07, 11.77)9%0.29Marital status4Married/cohabitingSingle1.63 (1.12, 2.36)0%0.79Maternal employment status5EmployedUnemployed1.24 (0.57, 2.70)73%0.01Maternal education6Primary or lessSecondary or more0.61 (0.36, 1.02)61%0.03Residence2RuralUrban0.84 (0.24, 2.91)0%0.59Mode of delivery2VaginalCesarean1.13 (0.28, 4.61)0%0.63Place of delivery2HomeHealth facility1.37 (0.00, 160,000)89%0.002Number of antenatal visits6InadequateAdequate0.72 (0.35, 1.47)76%0.006Antenatal breastfeeding counselling2YesNo1.93 (0.50, 7.43)0%0.56Knowledge of EBF4InadequateAdequate0.37 (0.10, 1.44)91% < 0.01Knowledge of duration of EBF2KnowledgeableNot knowledgeable1.90 (0.00, 1954)83%0.02Infant sex3MaleFemale1.08 (0.65, 1.78)0%0.47 Summary of pooled ORs for determinants of EBF In the subgroup analysis of the HIV positive mothers, only 3 determinants were reported in at least 2 studies: antenatal visits, maternal education level and employment status. The only significant determinant was employment status which had an OR 1.37 (1.10,1.72) favouring employed mothers (Fig. 3). Forest plot showing pooled or for employment status among HIV positive mothers Discussion This systematic review and meta-analysis of the published studies on the determinants of exclusive breastfeeding (EBF) among infants under six months in Tanzania highlights the factors influencing EBF and their implications for policy and practice. To the best of our knowledge, this is the first systematic review on the subject in Tanzania. The overall meta-analysis identified marital status as the only statistically significant determinant of EBF, with married or cohabiting mothers having higher odds of practicing EBF compared to single mothers. In the subgroup analysis of HIV positive mothers, employment status was identified as being positively associated with EBF. Our findings align with a meta-analysis from Ethiopia which also found that marital status was positively associated with EBF practice [10]. The higher rates of EBF in this group could be explained by spousal/partner support for the breastfeeding mother; both emotional and financial. Single mothers may have to focus on earning a living and hence may not be able to sustain EBF for longer periods. A systematic review of qualitative studies from Sub-Saharan Africa, described support from the husband as important to successful breastfeeding. This support could be economic through ensuring that mothers had enough food, and emotional through the spouse being faithful as unfaithfulness was reported as causing stress and worry for a breastfeeding mother, especially during the HIV pandemic [28]. However, other factors influencing EBF practices reported elsewhere including maternal education, antenatal care visits, antenatal counselling and employment status were not observed to be statistically significant in our pooled analysis, despite showing trends in individual studies. While higher education was positively associated with EBF in some Tanzanian studies [11, 21] this effect was not observed in the pooled meta-analysis. Meta-analysis from Ethiopia have previously reported higher EBF rates among women with a higher level of education compared to those with lower levels [10, 29]. This differs from evidence from China which showed lower EBF rates amongst women with higher levels of education [30]. The impact of maternal education on breastfeeding practices may be mediated by intermediate determinants of employment, financial ability and improved awareness. Women with higher education are more likely to be employed and this may negatively impact breastfeeding. Additionally, these women may have access and ability to purchase breast milk substitutes further influencing poor breastfeeding practices. Conversely however, higher levels of education may also imply improved awareness of the benefits of EBF, more access to information and improved health seeking behaviour all of which can positively influence EBF. Despite the UNICEF Baby Friendly Hospital Initiative (BFHI) emphasis on antenatal breastfeeding counselling as an important step towards breastfeeding success [31], its impact was inconsistent across the studies in our review; similarly, the number of antenatal visits had no association with EBF. This could possibly be due to variability in the content of the counselling, the frequency and the method by which it was delivered. Tanzania does not have a standardised curriculum for antenatal breastfeeding education hence there may be lack of uniformity in what women are taught during antenatal visits. A previous meta-analysis that assessed the impact of interventions related to step 3 of BFHI found that antenatal interventions are effective at increasing breastfeeding initiation, duration, or exclusivity where they combine both education and interpersonal support and where women’s partners or family are involved [32]. This underscores the importance of targeting interventions not only at pregnant women but at the wider support systems. Although maternal employment has been reported previously as negatively impacting EBF in meta-analysis from Ethiopia and Bangladesh [33, 34], this relationship was not observed in our meta-analysis. This is an unexpected finding since previous studies from Tanzania have shown low EBF prevalence among employed mothers [35, 36]. Possible explanations could be the different ways in which employment is defined in studies; some included only formal employment while others included self-employment. EBF practices would be different in these 2 groups where self-employed mothers are more likely to have flexibility in breastfeeding their infants. Maternal knowledge related to breastfeeding was not a significant determinant of EBF in our meta-analysis unlike a systematic review from China [37]. Overall, the knowledge of exclusive breastfeeding and its benefits is high in Tanzania and the wider East African region, with many mothers well informed about the advantages of breastfeeding [38, 39]. However, good maternal knowledge may not always translate into sound practices because infant feeding decisions in many African cultures are made by fathers or grandparents rather than mothers. Having experienced breastfeeding problems was a determinant reported in only 2 studies [13, 21] and 1 of them did not have raw data hence could not be pooled into the meta-analysis. Although not frequently reported in literature, breastfeeding problems such as pain, engorgement and low milk supply are common and may result in early cessation of EBF. A qualitative study from Tanzania found that perceived low milk supply was reported as barrier of EBF [40]. Lack of trained human resource to provide timely and relevant lactation support further compounds the problem. Notably, four studies out of the eleven in our meta-analysis focused on HIV-positive mothers [16–18, 21]. Employment was found to positively influence EBF. HIV positive mothers who were employed more likely to practice EBF as compared to those unemployed. This is in contrast to a study from South Africa that showed among HIV positive women on antiretroviral treatment, those who were unemployed were more likely to breastfeed. It is plausible that our findings are due to factors like higher levels of education or awareness among employed HIV positive women. Limitation The differences in objectives and methodology among the studies reviewed, the variations in data collection methods, the differences in study populations, sample size and the definitions of EBF and its determinants may have contributed to significant heterogeneity. The study populations were not representative of all regions of the country hence study findings cannot be generalized. Conclusions and recommendations This is the first study to synthesize evidence on determinants of EBF in Tanzania. Of the 11 studies reviewed, data from only 7 could be included in the meta-analysis. Most studies were from the Northern region of the country with minimal data from other areas. This demonstrates that the determinants of EBF have not been fully studied among the Tanzanian population. Our findings highlight that single women in Tanzania are less likely to exclusively breastfeed their infants and among the HIV positive women, being unemployed results in lower breastfeeding rates. For Tanzania to achieve global breastfeeding targets, we recommend that structured breastfeeding support programs should be developed and should prioritize single mothers and strengthen community support networks. Antenatal clinics should adopt uniform, culturally adapted EBF messaging to address knowledge and practice gaps. Among women who are HIV positive, ongoing support is required to ensure their infants receive optimal feeding. Future research is warranted in other geographical areas where practices and determinants may differ based on cultural norms. Qualitative approaches may be useful to explore unmeasured barriers like workplace factors or cultural practices. Longitudinal studies to assess causal relationships would strengthen the evidence base. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Acknowledgements The authors acknowledge the support received from the staff of the AKU Institute of Global Health and Development (IGHD) and Coletha Paradiso from Research office at AKU Medical College in Dar es Salaam. Abbreviations Baby Friendly Hospital Initiative Exclusive Breastfeeding Joanna Briggs Institute Human Immunodeficiency Virus Millennium Development Goals Odds Ratios Sustainable Development Goals Social Ecological Model of Health Tanzania Demographic and Health Survey Author contributions MN, MA, JD and ZB were involved in study conceptualization and planning. MN and RO developed the search strategy, retrieved articles, did data extraction and analysis. JD supervised the data extraction, analysis and synthesis. All authors have reviewed the final manuscript. Funding Funding for this study was provided through a fellowship seed grant from the Aga Khan University Institute of Global Health and Development. The funders had no role in the review process. Data availability This was a systematic review and all data is available from the manuscripts of the included studies and provided within the manuscript and additional files Declarations Ethics approval and consent to participate Ethics approval for this study was obtained from the Aga Khan University Ethics review committee (AKU-ERC), Ref no: AKU/2023/03/fb/011/018 as well as the National Institute of Medical research (NIMR), Ref no: NIMR/HQ/R.8a/Vol.IX/4626. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Data Availability Statement This was a systematic review and all data is available from the manuscripts of the included studies and provided within the manuscript and additional files