Trends and factors associated with early initiation of breastfeeding in Namibia: analysis of the Demographic and Health Surveys 2000–2013

Background Early initiation of breastfeeding (EIBF) lowers the risk for all-cause mortality in babies, including those with low birth weight. However, rates of neonatal mortality and delayed initiation of breastfeeding remain high in most low- and middle-income countries. This study aimed to assess the trends and factors associated with EIBF in Namibia from 2000 to 2013. Methods An analysis of EIBF trends was conducted using data from three Namibia Demographic Health Surveys. The present sample included singleton children younger than 2-years from 2000 (n = 1655), 2006–2007 (n = 2152) and 2013 (n = 2062) surveys. Descriptive statistics were used to analyse respondents’ demographic, socioeconomic and obstetric characteristics. Factors associated with EIBF were assessed using univariate analysis and further evaluated using multivariable logistic regression analysis. Results EIBF significantly decreased from 82.5% (confidence interval [CI]: 79.5–85.0) in 2000 to 74.9% (72.5–77.2) in 2013. Factors associated with EIBF in 2000 were urban residence (adjusted odds ratio 0.58, 95% CI: 0.36–0.93), poorer household wealth index (1.82, 1.05–3.17), lack of antenatal care (0.14, 0.03–0.81), small birth size (0.38, 0.24–0.63) and large birth size (0.51, 0.37–0.79). In 2013, factors associated with EIBF were maternal age of 15–19 years (2.28, 1.22–4.24), vaginal delivery (2.74, 1.90–3.93), married mothers (1.57, 1.16–2.14), delivery assistance from health professionals (3.67, 1.23–10.9) and birth order of fourth or above (1.52, 1.03–2.26). Conclusions Namibia has experienced a declining trend in EIBF rates from 2000 to 2013. Factors associated with EIBF differed between 2000 and 2013. The present findings highlight the importance of continued commitment to addressing neonatal health challenges and strengthening implementation of interventions to increase EIBF in Namibia. Electronic supplementary material The online version of this article (10.1186/s12884-018-1811-4) contains supplementary material, which is available to authorized users.

In the early 1990s, the WHO and United Nations Children's Fund (UNICEF) launched the Baby Mother-Friendly Initiative (BMFI) to support breastfeeding practices. Namibia adopted this initiative in 1992 and was among the first African countries to launch BMFI [16]. In Namibia, BMFI resulted in training healthcare professionals on breastfeeding management and promotion, certification of all 35 hospitals as baby-friendly [17] and an increase in EIBF from 52% in 1992 to 81% in 2000 [18]. However, in recent years, Namibia has experienced changes that have posed threats to these gains. The prevalence of human immunodeficiency virus (HIV) among pregnant women receiving antenatal care increased from 4.2% in 1992 to 19.3% in 2000, with breastfeeding contributing 30-40% of mother-to-child transmission of HIV [19]. Namibia was also reclassified as an upper middle-income country in 2009, but rates of unemployment and poverty remain high. To date, no studies in Namibia have investigated the potential effects of these changes on child health indicators, including EIBF.
Namibia also failed to meet targets for child health specified in the Millennium Development Goals (MDGs) [20]. Recent UNICEF estimates indicate the EIBF rate declined from 81% in 2000 to 71% in 2016 [1,20]. This downward trend and the lack of evidence on changes over time in factors associated with EIBF necessitate further investigations using nationally representative data. This study aimed to assess trends and factors associated with EIBF in Namibia from 2000 to 2013.

Data sources and sample
This study used nationally representative child datasets from the Namibia Demographic and Health Surveys (NDHS) for 2000, 2006-2007 and 2013 [18,20,21]. All surveys used a two-stage stratified cluster sampling design based on administrative regions and locations [18,20,21]. The first stage involved identification of primary sampling units and the second involved selection of households. Both stages were based on the sampling frame used in the Namibia Population and Housing Census preceding the NDHS (1991, 2001 and 2011). Individual households were selected using systematic sampling [18,20,21]. A trained team of interviewers using standardised pre-tested household, women's and men's questionnaires (translated into six local languages) collected data for all surveys [18,20,21].
This study used data for households and women aged 15 Figure S1). The present analysis included children younger than 24 months who were singleton live births (2000, n = 1655; 2006-2007, n = 2152; 2013, n = 2062). Detailed information on NDHS data sources, survey settings and sampling strategies have been described elsewhere [18,20,21].

Outcome variable
The main study outcome was EIBF, which was assessed among children younger than 24 months. EIBF is defined as putting a newborn baby to the breast within 1 h of birth [22]. The NDHS assessed EIBF by asking respondents, 'How long after birth did you first put (last born child's name) to the breast?' [18,20,21]. Responses were categorised into those who started breastfeeding within 1-h of birth and more than 1-h after birth.

Statistical analysis
Data were checked for completeness and consistency. Statistical analyses on complete cases were performed with STATA version 13.1 excluding 111 (6.7%), 209 (9.7%) and 155 (7.5%) children who had missing data on the outcome variable for 2000, 2006-2007 and 2013 surveys, respectively. We used frequencies and percentages to report sample characteristics and EIBF trends, and chi-square tests to assess associations between explanatory variables and EIBF in each survey. Simple and hierarchical multivariable logistic regression analyses were conducted to assess factors associated with EIBF in each survey. To allow comparability across the surveys, variables in all three surveys with a p-value > 0.25 in univariable regression analyses were excluded from the multivariable analyses [30]. Maternal, obstetric, and child-related factors were first included in the model separately and then together in a final model. We also adjusted for the sampling weight and cluster design of the surveys [31] and reported unadjusted and adjusted odds ratios (OR) with 95% confidence intervals (CI). Table 1 Table 2 illustrates EIBF rates by mother and child characteristics. In all three surveys, there were equal proportions of EIBF among male and female babies, and EIBF rates were higher among mothers who lived in urban areas, delivered vaginally and in health facilities and had four or more ANC visits. EIBF was significantly associated with birth size (2000 and 2006-2007), delivery mode (2006-2007 and 2013), and occupation and marital status (2013).

Respondents' characteristics
Overall, there was a significant decrease in the proportion of mothers who initiated breastfeeding early. The EIBF rate among urban mothers decreased significantly from 54.8% (95% CI: 50.9-58.7) in 2000 to 40 Table 2).

Factors associated with EIBF in Namibia
In the bivariate analysis, EIBF was significantly associated with birth order and birth size in 2000, birth size, maternal age, and delivery mode in 2006-2007 and birth size, birth order, delivery assistance by TBAs, delivery mode, ANC, occupation, wealth, education and marital status in 2013 ( Table 3).

Discussion
The WHO classifies EIBF rates as poor (0-29%), fair (30-49%), good (50-89%) and very good (90-100%) [2]. The EIBF rate in Namibia is still considered good despite the significant decline from 82.5% in 2000 to 74.9% in 2013. This trend is similar to those witnessed in other middle-income countries such as Vietnam (62%) and Haiti (69%) [29,32], and in lower-middle income settings (82%) [33][34][35]. The decline in the EIBF rate in Namibia between 2000 and 2006-2007 may be attributed to the high rates of HIV [19], insufficient health infrastructure, poor access, and ineffective and inefficient health service provision [36]. However, the slight increase in the EIBF rate between the 2006-2007 and 2013 surveys may be explained by government efforts, such as enactment of infant and young child feeding policies [36].
Our findings showed the number and nature of factors associated with EIBF varied from 2000 to 2013. Overall, this may be attributable to sociocultural and economic changes, increased rural-urban migration, improved school enrolment among girls, reduced teenage pregnancies, increased employment among women and health service use [37]. We found significant associations between EIBF and delivery mode, ANC attendance, baby's birth size, place of residence, maternal age, marital status, delivery assistance and birth order. EIBF was also more likely among mothers who had a vaginal delivery compared with a caesarean section [13,33,[38][39][40]. This may be partly explained by the increased rates of caesarean sections in 2006-2007 and 2013, effects of anaesthesia delaying the onset of lactation, associated respiratory distress among babies delivered by caesarean section [15] and healthcare professionals' increased preoccupation with assisting mothers to stabilise rather than initiating breastfeeding [11,41]. This highlights the need for   [33,42]. It is paramount to promote skilled birth attendance and baby-friendly initiatives in health facilities [22] and improve new mothers' breastfeeding practices through nutrition education during ANC visits [43]. Findings from Nigeria and Brazil indicate EIBF was more likely among mothers who had large babies at birth [9,14,25,26]. In contrast, we found that EIBF was less likely among both large-and small-sized babies between 2000 and 2006-2007 compared with average-sized babies. Small babies often have weak breastfeeding reflexes, poor coordination, and difficulty swallowing [14,26]. This may be attributable to healthcare providers' increased attention to stabilising the baby rather than easing the initiation of breastfeeding [26]. Existing literature also shows that both mothers and healthcare providers perceive large babies as healthy, leading to EIBF [26].
We found EIBF was more likely among urban mothers compared with rural mothers, which was consistent with previous studies [14,44]. This may be explained by higher ANC attendance rates, increased levels of employment and higher education levels among mothers living in urban areas. Urban women may also have increased access to information, leading to higher EIBF rates [45]. However, in 2013, place of residence was no longer a significant factor. This may be because of increased urbanisation and service provision to various parts of the country [20].
Younger women and adolescents had increased odds of initiating breastfeeding within the one-hour post-delivery period, which was similar to findings from low and middle-income countries [10,46]. This may be attributable to improved girls education, numbers of planned pregnancies and social support [32], and the intention to breastfeed and better prenatal attitude [47,48]. Moreover, maternal age as a determinant of EIBF is largely dependent on the presence of factors such education level and residency; in the absence of those factors, age may not impact the EIBF rate [10]. Other factors associated with EIBF included marital status and birth order. EIBF was more likely among married mothers and babies born into large families, which may be because of psychosocial support from family [49]. Multiparous women also have an increased level of knowledge and experience, and EIBF may, therefore, be more likely.
We did not find a significant association between socioeconomic status and EIBF, which was consistent with a study on trends and determinants of EIBF in Vietnam [29,44]. Education, occupation, and wealth were not significantly associated with EIBF, except in 2000 where women in the poorer quintile had increased odds of EIBF. This finding may show the diminishing influence of socioeconomic factors on the uptake of health Percentage point difference between survey years with a significance tests for difference in proportions; *p-value < 0.05; **p-value < 0.01 services and health information in Namibia. However, our finding differs from reports from Ethiopia [13] and Ireland [50] that showed EIBF was more likely among employed women, and from Indonesia [12] where it was less likely among women from wealthy households.

Strengths and limitations
The use of publicly available, nationally representative data in this study allows our findings to be generalised to Namibia. However, our study has some limitations and caution is needed in interpreting the results. First, data collection for our main outcome relied on respondents' recall, meaning there is a likelihood of recall bias. Existing literature shows that overestimation or underestimation of EIBF is possible because of the mothers' inability to assess time in minutes or hours [51]. Moreover, it has been found that a mother's response to the question on when they first put their baby to the breast 'was related to the first time the newborn received breast milk rather than their first attempt to initiate breastfeeding' [51]. Second, data on delivery mode were not available before 2006-2007. Thus, we could not assess the trends and association of delivery mode and EIBF in 2000, but the later years showed a consistent pattern. Lastly, causality cannot be inferred as this was a crosssectional study.
Policy and practice implications UNICEF and WHO are implementing a global initiative to improve breastfeeding outcomes with a goal of improving the average EIBF rate to 70% globally [1]. Namibia has achieved this target because of government commitment through the implementation of policies and programmes (e.g. Food and Nutrition Policy for Namibia, National Policy on Infant and Young Child Feeding, Food and Nutrition Guidelines) and a focus on accelerating the achievement of better child health indicators since 1993 [52]. However, the EIBF rate has declined over the recent years, eroding the gains of various programmes and policies. There is a need for increased focus on reviewing existing breastfeeding policies and ensuring full implementation of relevant breastfeeding policies and programmes such as BMFI to accelerate progress towards reversing this trend of declining EIBF in Namibia and contribute to achieving the sustainable development goals 3 on health.

Conclusion
Namibia experienced a declining trend in the EIBF rate from 2000 to 2013. Factors associated with EIBF also changed over the years. In 2000, urban residence, poorer women, ANC attendance and baby's birth size were associated with EIBF. Associated factors in 2013 were maternal age, marital status, caesarean section, TBA-assisted delivery, and birth order. These findings suggest there is a need for renewed commitment to promote breastfeeding in Namibia to reverse the trend of declining EIBF.

Additional file
Additional file 1: Figure S1. Flow diagram showing how the sample was obtained, provides information on the survey years, the number of households surveyed and the response rate both at the household level and for the women, detailing the number of births in the preceding 5 years, babies aged < 24 months old and the singleton babies. (DOCX 69 kb)

Availability of data and materials
The NDHS data and materials used in this study are available for free and on request on the Demographic and Health Survey website at www.dhsprogram.com. Delivery mode was not available for the 2000 NDHS. However, this was measured in the later surveys AOR adjusted odds ratio, CI confidence interval, COR crude odds ratio (unadjusted odds ratio), NDHS Namibia Demographic Health Survey, TBAs traditional birth attendants **p-value < 0.001 *p-value < 0.05