In our study on Nias, we found that more than half of the mothers (52%) initiated breastfeeding (la femenu) in the first hour after delivery. In total, 72% of the mothers breastfed their newborns in the first six hours after baby's birth This finding is in line with previous studies in Timor Leste, India, and Turkey [16–18]. However, 28% of the Nias mothers initiated breastfeeding six hours after birth or even later, and 17% discarded the colostrum due to local traditional beliefs similar to the ones in the study in India [19, 20]. Reasons given for not feeding colostrum to newborns included the traditional belief that colostrum was "dirty (n = 24), cheesy (n = 14), and indigestible (n = 8)" and that "children will suffer from stomach ache/afökhö talu (n = 21)", "children will get illnesses such as fever/mofa'aukhu (n = 15)", and "children will be stupid/bodo (n = 4)".
The notion that "colostrum is dirty" (n = 24) is likely due to the yellowish color of colostrum that is much different from the whitish color of mature mother's milk. The yellow color of colostrum was identified with 'dirty (ta'unö)' breast milk, while the white color of mature milk was believed to be 'clean' maternal milk. Colostrum was considered by some mothers to be 'cheesy (oŵoyu)' breast milk that had been produced late in the mother's pregnancy (n = 14) and had no nutritional value (n = 10). Colostrum was also believed to have a bitter taste/afeto (n = 12). Thus several mothers considered colostrum to be an inappropriate food for the newborn and believed it would be harmful to the infant's health.
"According to my mother-in-law and other senior female relatives, colostrum is not a healthy food for the newborns. This first milk is dirty and cheesy. If it is given to the newborn, my infant will suffer from stomach ache . . ." (Mother in Bawadesolo)
The practice of discarding colostrum is strongly rooted in the study area. We argue that the strong role of senior family members, particularly senior women such as the paternal grandmother, in prohibiting the consumption of colostrum for newborns has sustained this custom, and that it has been carried out for generations in the study area.
It is known that early initiation of breastfeeding will result in better establishment of breastfeeding practices, including longer duration of breastfeeding. This practice also ensures that newborns receive food with high nutritional value at the beginning of their lives. Early breastfeeding initiation can protect the newborn from potentially harmful pre-lacteal feeding practices which deprive the infant of the important immune-protective properties of colostrum .
Practice of pre-lacteal feeding and use of supplementary liquids
Wide use of pre-lacteal foods has been identified from previous studies in India, Ethiopia, Bangladesh, and Tanzania [17, 20–24]. The practice of pre-lacteal feeding was also common in our study area (personal information with Health and Nutrition officers CWS). Sugar water (n = 37) and infant formula (n = 25) were the two most preferred pre-lacteal liquids mentioned by the mothers. Several mothers believed that providing pre-lacteal liquids could help the newborns to resist hunger (n = 30). They perceived that because of the process of childbirth, the newborns experience fatigue, which then leads to feelings of hunger. Therefore, supplementary feeding serves the perception that this practice reduces infants' hunger in the first hours of their life.
Although colostrum is enough to sustain the nutritional needs of newborn babies without any additional foods , this concept was not widely known in the study area. Giving pre-lacteal foods (e.g. formula milk or anything else other than breast milk) delays the production of breast milk [25, 26]. If infants consume pre-lacteal foods or supplementary foods, this interferes with suckling at the breast, the release of prolactin and, ultimately, the production of the mother's milk . This leads to lower breast milk production and a shorter period of breastfeeding [27, 28]. Pre-lacteal foods are often a source of newborn infections and diarrhea [12, 17]. The delay of breastfeeding initiation and giving pre-lacteal feeds creates a vicious cycle. The practice of pre-lacteal feeding delays the initiation of breastfeeding and the delay in initiating breastfeeding promotes further pre-lacteal feeding.
This Nias study found that three out of four children received supplementary feeding (liquids) in the first seven days of their lives. The common reason for giving supplementary feeding in the study area was perceived breast milk insufficiency (n = 22), which was likely due to the delay of breastfeeding initiation. Bearing in mind that the frequency and the duration of breastfeeding are important factors in stimulating breast milk production, it is not surprising that incorrect breastfeeding practice such as providing supplementary foods results in insufficient production of breast milk.
Introduction of complementary foods (solid, semi-solid or soft foods)
The appropriate time for introducing complementary foods and the types of foods are crucial factors to be considered in appropriate infant and child feeding practices. According to Agudo et al. , findings from several studies showed that the majority of mothers in developing countries initiated complementary feeding too early. Other studies have had similar findings [29–31]. Our study also revealed that early introduction of solid and semi-solid foods was common in the study area.
Most of our respondents preferred to introduce complementary foods (solid, semi-solid, or soft food) when the children were one to four months old. This was in line with the results of our qualitative assessments that found that the majority of mothers believed that an appropriate time for introducing complementary foods was between one and four months (n = 32).
A perceived decline in the production of breast milk was the main reason mentioned for early introduction of complementary foods in the study area (n = 30). Practice of pre-lacteal feeding, skepticism about the sufficiency of their own breast milk production, inappropriate time of introduction of complementary foods, as well as infrequent and too short duration of breastfeeds, were apparently responsible for the decline in the maternal milk supply. On the other hand, perceived insufficient breast milk production also led to too early initiation of complementary feeding. In both situations, mothers believed that their child's crying was a signal of insufficient food intake. Thus, the provision of complementary foods seemed to them to be the only way to satisfy the child.
"Mothers in this area have usually introduced complementary foods when infants are four months old. However, they will provide complementary foods earlier if the child still cries although breast milk has already been given" (Mother in Tugala Gawu)
The opinions of senior female members of the family, especially the paternal grandmother (ina matua), about the infant's well-being played an important role in the decision of mothers to introduce solid, semi-solid, or soft foods (n = 23). If an infant was considered by the female in-law to be too 'thin' (afuo), a mother was likely to initiate complementary foods. As reported by respondents, the decision to provide complementary foods was usually not accompanied by maintaining breastfeeding frequency. It is therefore not surprising that increasing supplemental food was associated with less breastfeeding and breast milk production declined. Even worse, breastfeeding was generally no longer practiced after a child was given complementary foods. This is a very common pattern around the world reflecting a misunderstanding of complementary feeding (solids) as a replacement for breastfeeding, rather than a complement to it.
Early introduction of liquids and solids is unnecessary and unwise because it can reduce the duration and frequency of breastfeeding [12, 24], as previously observed in our study area. Premature introduction of complementary foods increases the risk of infant morbidity and mortality [12, 24]. We therefore conclude that children in our Nias study area face increased and unnecessary health risks due to inappropriate infant feeding practices.
The use of infant formula
Commercially produced infant formula was provided as a supplementary food (liquid) in addition to breast milk during the first seven days of life (32%), as a supplementary food in addition to breastfeeding within six months (13%), or to fully replace breast milk (2%) in the study area. Some mothers perceived that infant formula had highly valuable nutrients that were as good as, or even better than breast milk (n = 10). Therefore, they believed that if the household had the financial means, formula should be provided to the babies. However, according to Gribble et al.  and Mulder-Baalbergen et al. , several considerations should be taken into account when providing infant formula at the household level, particularly in emergency situations. These include the availability of clean water, energy for heating/boiling water and sterilizing bottles and other equipment, household finances as well as the ability of caregivers to prepare infant formula correctly according to the manufacturer's instruction [32, 33]. These requirements appeared to be very difficult for respondents in our study area to meet.
As stated in WHO/FAO guidelines , preparing infant formula according to the manufacturer's instructions is still not sufficient and safe because of the inherent bacterial contamination in many powdered formulas. Manufacturers do not usually give accurate information about an adequate temperature for the water to be mixed with the powdered formula. In order to destroy the bacteria, water must be boiled and then cooled to no less than 70°c (between 70° and 90°C) before mixing. This was also not known in the study area. Accurate assessment of water temperature is also highly uncertain.
Exclusive breastfeeding practice
In our study area, only one out of nine mothers practiced exclusive breastfeeding with their infants. This low prevalence of exclusive breastfeeding has been commonly reported in studies in India, Bangladesh, Sri Lanka, Turkey, and Tanzania [18, 20, 23, 24, 30]. Based on qualitative assessment, the most common barrier to practicing exclusive breastfeeding in the study area was perceived insufficient breast milk production (n = 30). Other reasons, such as believing that the child was starving (n = 15), the mother's activity outside her home (n = 10), poor knowledge of the benefits of exclusive breastfeeding (n = 7), and the mother's illness (n = 7) were also mentioned by several mothers.
A mother may make a decision to add supplemental fluids or soft complementary foods at too early an age if she is feeling discouraged about her milk production. If a mother believes that she cannot produce a sufficient quantity of breast milk, she tends to decrease her breastfeeding frequency. This will result in decreasing breast milk production.
"I am sure that I did not have sufficient breast milk production. My breast is too small and I am too thin as well. My mother-in-law said that this child (pointed out her two-year old son), was too thin when he was an infant, although I had already given my breast milk to him. Therefore she said to me that I should give him other foods as he was three months old." (Mother in Tögideu)
Infrequent and brief feeds are commonly practiced among mothers in the study area (personal communication with Health and Nutrition Officers CWS Nias). These two factors combined were likely to be main contributors to low breast milk production. This 'insufficient breast milk syndrome' has led to the belief that mother's milk could not provide sufficient nutrients for the infants. Thus, most of the mothers in the study area believed that a six month period of exclusive breastfeeding could endanger the health and nutritional status of their young children. Indeed, this would be the case, if babies were not fed frequently enough.
Brief feeds are actually very common in some regions where mothers are constantly active doing manual agricultural work or gathering water or fuel. When these mothers carry their babies with them (mostly on their backs), they may compensate for the short feeds by a more frequent feeding pattern.
In addition to the constraints identified above, we surmise that nutrition education on Nias Island might not be optimal with regard to promoting appropriate breastfeeding practices, and might be contributing to the low prevalence of exclusive breastfeeding in the first six months in our study area. Based on Agampodi et al. , a major determinant of exclusive breastfeeding practices is the health care provider's knowledge about, attitudes towards, and skills for promoting exclusive breastfeeding. Therefore, we speculate that insufficient breastfeeding-related knowledge among health personnel and community workers may have reinforced mothers' and families' perceptions that have led to the low prevalence of exclusive breastfeeding rates in our study area. With reference to communication to Health and Nutrition Officers of CWS Nias, it may, for example, be that the policy change of 2004 that adjusted the recommendations for exclusive breastfeeding from "four months" to "six months" had not reached the grass roots level of public health staff on Nias Island - even almost a decade after this policy change.
Breastfeeding prevalence at admission
Breastfeeding prevalence among respondents was low at admission into the study program for mildly wasted children on Nias Island. One of the reasons was probably related to the high proportion of children who were ≥ 24 months (75%). Another reason was presumably related to the heavy workloads of mothers in the study area. We found that women in this study had to perform income-generating activities for an average of six hours per day in informal sectors. Most of these activities took place away from their homes. Therefore, it would likely be difficult to breastfeed during their absence from home. This was usually followed by a decline of breastfeeding frequency and, as previously stated, a decrease in breast milk production and, ultimately, cessation of breastfeeding. However, such constraints can be prevented if the mothers have proper information on how to maintain milk production while separated from the baby, and how to compensate for the absence with more frequent breastfeeds when mother and child are together. A six hour separation is not an insurmountable barrier to exclusive breastfeeding.
We also found several cultural factors influencing the decision to cease breastfeeding. In our study area, the perception that prolonged breastfeeding would interfere with the child's growth is widespread. Several mothers believed that children who breastfed until the age of two years or beyond would favor consuming breast milk instead of family foods (n = 23). They believed that the child's daily nutrient intake would not meet the recommended levels if consumption of home-based family food was too low, and this would lead to the impaired growth of children (personal communications with several caregivers). We argue that the concern that children will not eat family foods properly if they are still breastfed is essentially a child-rearing issue. Parents can and should set reasonable limits on fulfilling children's consumption wishes beyond infancy. Parents need to understand that the way feeding is organized is a management question, and for optimal infant feeding the consumption of family foods should take place alongside of continued breastfeeding during a child's first two years.
The belief that pregnant women should not breastfeed their children was also a factor hindering the practice of breastfeeding during pregnancy (n = 15). According to Nias culture, breastfeeding by a pregnant mother could harm the fetus because it would decrease the fetus' food intake (personal communication with traditional healers in Sirombu, Nias). They also claimed that a pregnant mother who persisted in breastfeeding would deliver a thin (afuo), sickly (mofökhö-fökhö), and feeble-minded (bodo) newborn baby. Therefore, most of the pregnant mothers in our study area did not breastfeed their young children. Considering that short birth spacing was commonly found in the study area, a new pregnancy would lead to a decision to wean too early. This probably contributed to the low breastfeeding rate in our study.
The belief that a mother who is ill should not breastfeed her child was widespread in the study area. According to local beliefs, the illness of a mother is transmitted to the breastfed child through breast milk. Therefore, some mothers in our study area did not breastfeed their children when they were ill (n = 9). In this situation, mothers preferred to provide family foods to the young children and halted breastfeeding during the illness. Unfortunately, the elimination of breastfeeding practices affects the breast milk production. A decreased breast milk supply was usually followed by breastfeeding cessation. This may also have contributed to the low breastfeeding prevalence in our study area.
The importance of appropriate educational programs
Bearing in mind that the protective effect of breastfeeding is especially significant in populations with high infant mortality, low literacy, poor sanitation facilities, poor nutritional status and generally low economic status , accurate information on the importance of early initiation of breastfeeding, six months of exclusive breastfeeding, and continued breastfeeding after the introduction of complementary foods would be particularly important for our respondents in this Nias study who likely faced similar living conditions. Therefore, the promotion of correct breastfeeding practices should be high on the agenda for any health-nutrition activities aimed to improve the well-being of young children on Nias Island.
It is also important to motivate mothers to practice exclusive breastfeeding for six months with continued breastfeeding until two years or beyond, as well as introducing timely complementary feeding (six months), since these are high-priority infant feeding indicators for child survival .
Appropriate infant feeding promotion would help prevent faulty feeding practices, such as the provision of pre-lacteal and supplementary foods (liquids) and too early introduction of complementary foods. These inappropriate infant feeding practices increase the risk of illness, malnutrition, and even death among infants and young children .
Special attention to appropriate breastfeeding and complementary feeding interventions should be given to the mothers/caregivers of young infants of the ages most likely to be affected by malnutrition. This susceptible period is when young infants are introduced to foods and liquids other than breast milk. Focusing educational programs on this target group is likely to be more cost-effective than interventions that include a wider range of mothers/caregivers .
Our qualitative assessments also found that the support of family members was an important influencing factor for improving infant feeding practices in the study area. We learned that paternal grandmothers have great influence on infant/child feeding decisions, such as duration of exclusive breastfeeding, and the time to introduce complementary feeding , as was also observed by Aubel in the Grandmother Project . Therefore any interventions aimed at improving young child feeding practices, particularly infant feeding behaviors, should also target paternal grandmothers or partners of pregnant women, and the families of newborn infants, especially those members who play an important role in caring for and feeding young children. Grandmothers, who were reported in our study to have a negative influence on breastfeeding, should be included in programs for breastfeeding promotion. Other target groups such as community workers, health professionals, and traditional birth attendants should also be given the necessary guidance, appropriate training, and support with respect to breastfeeding promotion.