The status of nutritional knowledge, attitude and practices associated with complementary feeding in a post‐conflict development phase setting: The case of Acholi sub‐region of Uganda

Abstract Inappropriate complementary feeding is an important challenge to proper child nutrition in post‐conflict rural areas in many sub‐Saharan African countries. While in protected areas during conflict situation and soon after during recovery, communities normally receive nutrition education as part of capacity building to improve knowledge, attitude, and practices to enable them manage maternal and child nutrition issues during the post‐conflict development phase. It is largely unknown whether capacity in nutrition provided is maintained and adequately applied in the post‐conflict development situation. Using Acholi sub‐region of Uganda, an area that experienced violent armed conflict for 20 years (mid‐80s–early 2000), as a case study, we examined the status of nutritional knowledge, attitude, and practices associated with complementary feeding among caregivers of 6‐ to 23‐month‐old children in a post‐conflict development phase following return to normalcy nearly 10 years post‐conflict emergency situation. The results showed that a high proportion of caregivers had good knowledge (88%) and attitude (90.1%) toward complementary feeding. However, only a half (50%) of them practiced correct nutrition behavior. Education status of the household head and sex of the child significantly predicted caregiver knowledge on complementary feeding (p ≤ 0.05). Education status of the household head also predicted caregiver attitude toward complementary feeding (p ≤ 0.05). Poverty, food insecurity, and maternal ill health were the major factors that hindered caregivers from practicing good complementary feeding behavior. These results demonstrate that nutrition education on complementary feeding provided to the community during conflict emergency and recovery situation is largely retained in terms of knowledge and attitude but poorly translated into good child feeding practices due to poverty, food insecurity, and maternal ill health. Maternal health, food security, and poverty reduction should be prioritized if adequate complementary feeding is to be achieved among conflict‐affected communities in the post‐conflict development phase.


| INTRODUC TI ON
Nutrition is a vital component of health promotion and disease prevention (Mowe, Bosaeus, & Højgaard, 2008). The impacts of nutrition on health throughout the course of human life are very profound and are inextricably linked to cognitive and social development, more so in early childhood (Black et al., 2013). The major effects of undernutrition are believed to occur during the first 2 years of human life. This is because, at this stage, undernutrition causes irreversible damage to physical, mental, and social development of the child transcending into reduced intellectual potential at adulthood (Motee, Ramasawmy, Pugo-gunsam, & Jeewon, 2013;Semahegn, Tesfaye, & Bogale, 2014;WHO, 2010b;World Bank, 2006). Despite the well-recognized importance of proper child nutrition to health well-being and human capital development, child undernutrition has remained one of the main public health problems in developing countries (Müller & Krawinkel, 2005;Semahegn et al., 2014). It is one of the most common causes of morbidity and mortality among children in those countries (Amsalu & Tigabu, 2008). It has been estimated that more than 1/3 of the under-five mortality in developing countries are due to undernutrition-related diseases (Daelmans et al., 2013;Mesfin, Henry, Girma, & Whiting, 2015). This situation has largely been attributed to inappropriate complementary feeding practices by caregivers who in most cases lack adequate nutrition knowledge and information (Khanal, Sauer, & Zhao, 2013;Shi & Zhang, 2011).
A appropriate diet is critical especially in the first 2 years of a child's life (Rao, Swathi, Unnikrishnan, & Hegde, 2011;World Bank, 2006). This period is considered the most crucial time due to the increased nutritional needs to support rapid growth and development (Semahegn et al., 2014). Infants and young children in developing countries are at an increased risk of undernutrition from 6 months of age onwards when complementary foods are introduced (Memon, Shaikh, Kousar, & Rubina, 2010;Muhimbula & Issa-zacharia, 2010;Rao et al., 2011). The susceptibility of young children to undernutrition becomes apparent if complementary foods are of low nutrient density and bioavailability (WHO, 2013). On the other hand, if complementary feeding is not carried out properly, it can lead to problems such as diarrhea, growth retardation leading to kwashiorkor, marasmus, and immunodeficiency marked by recurrent and persistent infections which may be fatal (Rao et al., 2011).
Nutritional knowledge, attitude, and practices of caregivers are critical elements that determine the outcomes of complementary feeding regime administered to children (Saha et al., 2008;Turyashemererwa, 2009). It is classically believed that good knowledge should be associated with good attitude and proper nutritional practices (Azizi, Aghaee, Ebrahimi, & Ranjbar, 2011;Mowe et al., 2008). However, in some situations, good knowledge and attitude do not necessarily translate into good practices (Bukusuba, Kikafunda, & Whitehead, 2010). Nonetheless, in order to achieve proper outcome of complementary feeding, it is essential that nutritional knowledge, attitude, and practices should be appropriate.
In the context of this study, knowledge is defined as the caregiver's understanding of complementary feeding and nutrition, including the ability to remember and recall food, nutrition and specific pieces of information, and facts essential to complementary feeding.
Attitude is defined as the emotional, motivational, perceptive, and cognitive belief that positively or negatively influences the behavior or practice of a caregiver toward complementary feeding. Practice is defined as the observable actions of the caregiver that could affect the nutrition of the child undergoing complementary feeding such as eating, feeding, washing hands, cooking, and selecting foods (FAO, 2014).
In many developing regions of the world, a number of countries have experienced armed conflict (Cortez & Kim, 2012;Gates & Strand, 2012) and many more are currently experiencing the same.
Depending on the nature of conflict, the affected communities are either placed in Internally Displaced Peoples (IDP) camps (e.g., Northern Uganda during the 20-year-old LRA armed conflict from mid-80s to mid-2000) within the country or in Refugee Camps (as is the case with South Sudanese refugees currently in several locations in Uganda). From a nutritional point of view, displaced people living in protected areas are typically confronted with two phases of practical adjustments to cope with. First, being immobilized in protected areas during the emergency situation, mechanisms to ensure proper nutrition especially among vulnerable groups such as women (especially pregnant and lactating) and children (under 5 years) need to be devised. As part of humanitarian assistance during emergency and post-emergency recovery phase, a number of agencies, especially Non-Governmental Organizations (NGOs), provide nutrition education and to some extent nutritious food products tailored to nutritional needs of vulnerable people. The second phase of adjustment is during the post-conflict (or post-emergency) development phase when conflict-affected communities return to original homes. In this situation, external humanitarian assistance diminishes and returnees are expected to live economically on their own. Therefore, the returnees should make use of nutrition education provided during the emergency and recovery phases to practice good nutrition behavior.
An important question which has largely remained unanswered is as to whether the capacity building in nutrition provided during emergency and recovery phases of conflict enables the affected communities to exhibit good nutritional knowledge, attitude, and practices during the post-conflict development phase. This question is important because, during the post-conflict development phase situation, affected communities are normally confronted with several challenges, and as such, the issue of proper and adequate nutrition of vulnerable people can easily be neglected. Using the Acholi sub-region of Uganda as a case study, we examined whether, during the post-conflict development situation, conflict-affected communities are able to exhibit appropriate nutritional knowledge, attitude, and practices required for a better outcome of complementary feeding of children aged 6-23 months. In Acholi sub-region, people lived for over a decade in IDP camps from mid-80s up to mid-2000 (Saile, Ertl, Neuner, & Catani, 2014) and several humanitarian agencies provided nutritional education to the affected people (ACF (Action Against Hunger), 2004). It is now about 10 years since people in the sub-region returned to their original villages following official closure of the IDP camps in 2006.

| Study design, area, and population
A cross-sectional study design involving in-depth household interviews and focus group discussions was applied. The study was conducted in Nwoya (02°38′N 32°00′E) and Amuru (02°50′N 33°05′E) districts of the Acholi sub-region. Amuru district is composed of 6 sub-counties, 32 parishes, and 67 villages while Nwoya district is composed of 5 sub-counties, 25 parishes, and 63 villages. The two districts were selected because they had the highest rates of undernutrition in the sub-region (WFP (world Food Programme), 2012) despite substantial nutrition education provided to the community during conflict emergency and recovery situation. The study population consisted of caregivers of children aged 6-23 months old. In the context of this study, a caregiver is the mother of a child or another person who takes care of the child in the absence of the mother.
The inclusion criterion was that the caregiver should have received nutritional training while living in the IDP camp and or immediately after camp closure in 2006.

| Sample size and sampling framework
The number of participants (n) that participated in the study was determined according to the method of Krejcie and Morgan (1970). Using the combined number of households (N = 315,678) for Amuru (188,860) and Nwoya (126,818) district (UBOS (Uganda Bureau Of Statistics), 2014), the sample size was determined to be 382 respondents. Household statistics were used to estimate the sample size because data on the number of caregivers are usually unavailable. To locate the study participants, a multi-stage sampling framework was applied. First, 2 sub-counties were randomly selected from each district. From the selected sub-counties, 3 parishes were randomly selected followed by random selection of 2 villages per parish. Random selection of sub-counties, parishes, and villages was applied because all the two districts were affected by the 20-year armed conflict. Finally, participants were purposively selected from each village taking into account the inclusion criterion indicated above.

| Study instruments
Knowledge, attitude, and practices (KAP) were assessed using a standard questionnaire adapted with modification from WHO (2010a) and FAO (2014). The questionnaire also had a provision for collection of data on socio-demographic and economic characteristics of the households. With regard to knowledge, questions provided for correct and wrong answers while for practice, caregivers were asked to provide information on how they conducted complementary feeding. For attitude, both positive and negatively framed questions were provided and the respondent had three response options to choose from (disagree, neither agree nor disagree, agree) according to Anand and Puri (2013). A focus group discussion (FGD) guide was used to generate information to supplement what was obtained from household interviews. The instruments were pretested among caregivers of children in the nutrition ward at Gulu Regional Referral Hospital, located in Northern Uganda. This was done to ensure accuracy, clarity, and consistency in the interpretation of questions. After pretesting, responses were analyzed to check for validity and ambiguous questions were rephrased.
Another round of pre-test exercise was conducted in Agwee village, Laroo Division, Gulu district, to check for reliability of the instruments using the Test-Retest method (Birkman, 2001).

| Data collection
Data were collected using research assistants who had previous exposure to nutrition surveys and fluent in English and local language used by all members of the community in the study area (Luo/ Acholi). The assistants were conversant with the instruments since they had participated in the pretesting exercise. The assistants were trained only to interpret questions to the study participants but not to assist them in providing answers. Focus group discussions (FGDs) were conducted prior to household interviews. Each FGD constituted of 6-10 members, and a total of 10 FGDs were conducted. This number of FGDs is above the minimum number of six that has been reported to ensure saturation of information in qualitative studies (Guest, Bunce, & Johnson, 2006;Mclafferty, 2004). To ensure originality of information, households that participated in the FGDs were not selected for household interviews. During the interviews with the caregivers, other members of the household were not allowed to participate to avoid influencing the caregiver.

| Data analysis
Data from FGDs were transcribed, coded, and common themes established and analyzed using qualitative content analysis method (Cho & Lee, 2014). With regard to knowledge and practices, every correct answer took on a score of 1 while a wrong answer and where the respondent did not know, a score of 0 was given. In the case of attitude, a score of 0, 1, and 2 was given to "disagree," "neither agree nor disagree," and "agree" response, respectively. However, reverse scoring was done for negatively framed statements. This meant that a score of 0 was given to "agree" and 2 to "disagree" (Anand & Puri, 2013). Scores on knowledge, attitude, and practices for each respondent were calculated by summing up the scores attained on each question and the overall score ranked as good or poor. Knowledge, attitude, and practice scores were ranked as poor if the overall score fell below 55%, 57.1%, and 62.5%, respectively, according to Ul Haq, Hassali, Shafie, Saleem, and Farooqui (2012).
Data on socio-demographic characteristics were summarized using descriptive statistics (frequency, percentages, mean, and standard deviation). Binary logistic regression was performed to determine socio-demographic predictors of good knowledge, attitude, and practices, respectively. The logistic regression model used is pre- where Y i is the binary dependent variable depicting the status of knowledge, attitude, or practices; X 1 to X 14 are the independent variables; α is the regression constant; β is the regression coefficient; and μ is the error term. Before running the regression, correlation analysis (Pearson) was run to eliminate highly correlated independent variables (correlation coefficient greater than 0.70) according to Dormann et al. (2013). The following independent variables were selected and used to run the binary logistic regression: sex of the child (male = 1, female = 0), education status of the household head (primary and above = 1, no formal education = 0), sex of the household head (male = 1, female = 0), age of (1) Data based on sample size of 382 caregivers. A caregiver is the mother of a child or another person who takes care of the child in the absence of the mother.
the caregiver (years), marital status of the caregiver (married = 1, otherwise = 0), employment status of the household head (formal = 1, non-formal = 0), employment status of the caregiver (formal = 1, non-formal = 0), education status of the household head (primary and above = 1, no formal education = 0), education status of the caregiver (primary and above = 1, no formal education = 0), size of the household, amount of money spent on food per month (UGX), residence of the caregiver (rural = 1, urban = 0), age of the child (months), birthplace of the child (health care center = 1, otherwise = 0), and attendance of nutrition training (yes = 1, no = 0).
Binary regression was run separately for each dependent variable. The binary dependent variable was a dummy for good or bad knowledge, attitude, and practices, respectively. Statistical analyses were performed using Statistical Package for Social Sciences (SPSS) version 20.

| RE SULTS
Data on socio-demographic characteristics of the study participants are presented in Table 1. In terms of age, the majority of the caregivers (90%) were between 35 and 45 years, 67.3% of the respondents resided in rural areas, while 84.0% were married. The number of male children in the households sampled was higher than the number of females by 5%. The mean age of the children was 13.8 months and ranged from 6 to 23 months. Majority of them (57.2%) were at 6 months (9.9%), 10 months (7.9), 11 months (9.2%), 12 months (7.9%), 14 months (7.9%), 19 months (7.2%), and 20 months (7.2%). More than ¾ of the infants were born from a health facility. In terms of leadership, the majority ( The distribution of scores on knowledge, attitude, and practices regarding complementary feeding among caregivers is presented in Figure 1. With regard to knowledge and attitude, the proportion of caregivers who had good nutrition knowledge and good attitude toward complementary feeding was above 80%. On the other hand, the proportion of caregivers with good practices on complementary feeding was identical to those who exhibited poor practices ( Figure 1).
Data on specific aspects used to test caregiver's knowledge on complementary feeding are presented in Table 3. The proportion of caregivers with correct responses was higher (above 60%) than those with wrong responses for all aspects tested with the exception on the aspect testing the importance of including animal foods in a child's diet.
Data on specific aspects of complementary feeding used to test attitude are presented in Table 4. Generally, caregivers' attitude toward complementary feeding was good and above 70% with the exception that more than half of them found breastfeeding for 2 or more years embarrassing.
Distribution of caregivers' adherence to recommended complementary feeding practices is presented in Table 5. Ninety percent of the respondents adhered to continued breastfeeding along with complementary foods. However, less than 50% of the respondents gave their children the right consistency (6 months old: thick mashed or semi-solid food; 8 months old: finger foods; and 12 months old: family food) of complementary foods, met the minimum meal frequency (2-3 and 3-4 times per day for breastfed children 6-8 and 9-23 months, respectively, and 3-4 times for non-breastfed 6-to 23-month-old children), introduced complementary foods at 6 months of age and ensured dietary diversity (consumption of food from at least 4 or more food groups).
Results of binary logistic regression analysis used to test the effect of socio-demographic factors on knowledge, attitude, and practices regarding complementary feeding are presented in Table 6. Only, the educational status of the household head and the sex of the child were significant predictors of knowledge on complementary feeding (p = 0.009 and 0.005, respectively). The relationship is presented in equation 2.
TA B L E 2 Community perspective on complementary feeding situation

Adequacy of complementary feeding
Generally, the community believes that it is adequate because caregivers should follow advice from the hospital. However, adequacy depends on the household and specific mothers. Some household lack food while some mothers may have health complications which bar them from following the recommended practices

Initiation of complementary feeding at 6 months
Varies from household to household. Majority of the caregivers start at 6 months while others start early (3 or 4 months) due to various health complications including inadequate breast milk and/or swollen breasts/breast engorgement Feeding at the recommended meal frequency (2-3 and 3-4 times for 6-to 8-month-old and 9-to 23-month-old children, respectively) Some caregivers adhere while others do not. Most caregivers do not adhere because of too much work/ responsibilities at home. Following the recommended meal frequency is also challenged by weather changes which affect crop yields and availability of food Hospitals, friends, women groups, grandmothers, mothers-in-law, village health teams, relatives, radios, own instinct from birth, and Non-Governmental Organizations.
Benefit and adequacy of information on complementary feeding from the sources named For a few households, the information has helped to keep children in a healthy state and prevent some of the nutrition-related diseases. But for most of the households, the information is inadequate Other aspects of complementary feeding that the community would be interested to receive information on Formulation and preparation of nutritious infant foods, hygiene, estimation of the right quantity for infant feeding, education about advantages of good nutrition, and the recommended meal frequency Perception about the effect of inappropriate complementary feeding on the nutritional status of children Caregivers appreciated that inappropriate complementary feeding is bad because some children in the community who are poorly fed become very weak and malnourished.
Challenges experienced by mothers in the community to implement good complementary feeding practices Main challenges included: lack of financial resources to procure nutritious foods for the children, thus most households feed children on the same food type for over a week (lack of dietary diversity); inability to breastfeed due to pain in the breasts/swollen breasts/engorgement; Changing weather pattern which affects the yields of different crops; HIV/AIDS; and limited availability of diverse food categories Suggestions to improve complementary feeding Increase accessibility of affordable complementary foods, trainings on formulation of nutritious foods from locally available food resources, need for a local formula for producing infant feeds, and training on aspects of sanitation, nutrition, and agriculture a Information provided by caregivers of children 6-23 months during Focus Group Discussions (FGDs); 10 FGDs were conducted. A caregiver is the mother of a child or another person who takes care of the child in the absence of the mother.
Where Y represents good knowledge on complementary feeding, X 1 is the dummy for the educational status of the household head (primary and above = 1; and no formal education = 0), and X 2 is the dummy for sex of the child (male = 1 and female = 0). By equation 2, an increase in the education level of the household head by one unit reduces the level of nutritional knowledge of the caregiver by a factor of 2.9. On the other hand, having a male child is associated with a 2.8-fold increase in the level of caregiver's nutritional knowledge.
With regard to attitude, educational status of the household head was the only significant predictor (p = 0.011) ( Table 6). The prediction is presented in equation 3.
where Y represents good attitude towards complementary feeding and X 1 is the dummy for the educational status of the household head (primary and above = 1; and no formal education = 0). According to equation 3, an increase in the educational level of the household head by one unit is associated with a 3.2-fold decrease in the status of the attitude of the caregiver toward complementary feeding. However, none of the socio-demographic factors significantly predicted good complementary feeding practices (p > 0.05).

| D ISCUSS I ON
Nutrition education is an important vehicle for achieving good nutrition and nutrition-related health outcomes (Negash et al., 2014).
When adequately provided, nutrition education should enable the recipient to attain reasonably good nutrition knowledge, have a positive attitude toward good nutrition, and above all exhibit good nutritional practices (Ickes et al., 2017;Inayati et al.,  (3) Y = −11.675 − 3.199X 1 F I G U R E 1 Distribution of knowledge, attitude, and practices on complementary feeding among caregivers. Data based on a sample size of 382 caregivers. A caregiver is the mother of a child or another person who takes care of the child in the absence of the mother  Data based on sample size of 382 caregivers. A caregiver is the mother of a child or another person who takes care of the child in the absence of the mother.
factors are favorable, it would be expected that a combination of good nutritional knowledge and positive attitude toward nutrition should be appropriately translated to different aspects of nutritional practices/food habits (Rezaee, Azizi, & Hoseini, 2012). Much of the understanding gained on the impact of nutrition education on nutritional knowledge, attitude, and practices has been derived from studies conducted among communities living in peaceful environments. However, little is known about how nutrition education provided to communities living under stressful conditions (e.g., refugee camps) impact on their nutritional knowledge, attitude, and practices when they return to peaceful conditions. This study therefore examined the lacuna with a specific focus on complementary nutrition of children 6-23 months using the Acholi sub-region of Uganda as a case study.
Interestingly, the results presented in Figure 1   Data based on sample size of 382 caregivers. A caregiver is the mother of a child or another person who takes care of the child in the absence of the mother. a The question was reverse coded.

Proportion of caregivers that adhered (n) (%)
Continued breastfeeding along with complementary foods 344 90.1 Nature of food/food consistency for 6-to 8-month-old, 9-to 11-month-old, and 12-to 23-month-old children 113 29.6 Minimum meal frequency for 6-8, 9-11 months and non-breastfed children Data based sample size of 382 caregivers. A caregiver is the mother of a child or another person who takes care of the child in the absence of the mother.
TA B L E 5 Distribution of caregivers' adherence to recommended complementary feeding practices 2012; Vardanjani et al., 2015). The low level of formal education attained by the respondents in the study area (Table 1) could partly be responsible for the low adherence to recommended complementary feeding practices observed (Figure 1). A closer look at Table 1 reveals that most of the caregivers (67.1%) had attained only primary education. Education affects employability potential (Wambugu, 2011). People who are unemployed find it difficult to afford some of the basic necessities and requirements for their children.
Poverty and food insecurity might also account for the poor status of complementary feeding practices. This was clearly revealed by the caregivers during FGDs (Table 2). Poverty has been reported in many other areas as a hindrance to complementary feeding. Victor, Baines, Agho, and Dibley (2014) attested to this in a report which showed that poor economic status was one of the main risk factors for inappropriate complementary feeding practices in Tanzania.
Dietary diversity is essential for achieving good nutrition outcome from complementary feeding (Moursi et al., 2008). Table 5 shows that only a small proportion (47.4%) of the caregivers adhered to the recommended dietary diversity level. Agriculture is the principal economic activity in Acholi sub-region of Uganda.
FGD exercise revealed that changing weather pattern led to poor crop yields, and food insecurity and poverty in turn (Table 2). In a broader sense therefore, limited adherence to the recommended dietary diversity in complementary feeding can in part be attributed to food insecurity and poverty occasioned by the inability of households to adapt to weather changes. The dependence of food and nutrition security on agriculture in rural areas in developing countries has long been recognized (Titus & Adetokunbo, 2007). Classically, food and nutrition security can be met through the market (after the sale of produce) and/or own production pathways (Baiphethi & Jacobs, 2009). However, as demonstrated in this study, crop failure as reported by the community during FGDs (Table 2) is a critical factor that requires attention at both policy and household practice levels.
Despite the fact that a high proportion of caregivers (80%) had good nutritional knowledge, there were certain critical aspects where they exhibited deficiency. For instance, very low proportion of caregivers were knowledgeable about the importance of including animal foods in the child's diet (Table 3). Whereas this situation could in part be attributed to the low level of education among the caregivers (Table 1), it is also possible that caregivers might not have received adequate information on this specific aspect of nutrition knowledge. This suggests that the various sources of information on complementary feeding revealed by caregivers during FGD (Table 2) may not be providing sufficient information on such a critical aspect.
On the other hand, based on results presented in β: regression coefficients. SE: standard error. A caregiver is the mother of a child or another person who takes care of the child in the absence of the mother.
as demanded by the caregivers during FGDs ( Education status of the household head and sex of the child were the major predictors of good nutritional knowledge (Table 6).
Based on the coefficient of the child sex element in equation 2, it is apparent that caregiver's knowledge on complementary feeding is 2.8 times better when the baby is a boy child compared to when it is a girl child. This is not surprising because in many developing countries, especially in rural settings, the boy child is always more treasured compared to the girl child in terms of resource allocation (Dercon & Singh, 2011;Mehtabul & Geeta, 2011). This observation draws credence from studies conducted in Ethiopia which showed that mothers who had baby boys timely initiated complementary feeding than those who had baby girls (Semahegn et al., 2014).
The negative effect of the presence of a girl child on nutritional knowledge of the caregiver can compromise the quality of complementary feeding regime offered the girl child and nutritional outcomes as well. Strategic efforts are required to deal with the identified child sex disparity in complementary feeding. With respect to education status element of equation 2, it is interesting to note that an increase in education status of the household head is associated with a 2.9-fold decrease in caregiver's knowledge on complementary feeding. Education status of the household head also affected caregiver's attitude, but negatively. Looking at the coefficient of equation 3, it is evident that an increase in the education level of the household head is associated with a 3.2-fold decrease in attitude toward complementary feeding. Education is associated with employability, a factor which affects the length of time of interaction between the educated household heads and the caregivers (Wambugu, 2011). In the context of this study, it can be asserted that educated household heads are easily employed and spend most of their time at work, and as such, they are unable to adequately share knowledge on complementary feeding with the caregivers. Therefore, since the caregivers are unable to receive adequate information from the household heads, their attitude toward complementary feeding is negatively affected in the long run.
A major limitation of this study is that the community in the study area could have received some nutrition education which influenced the results. Future studies should control for this scenario.

| CON CLUS IONS
The results of this study have demonstrated that nutrition education on complementary feeding provided to the community during conflict emergency and recovery situation is largely retained in terms of knowledge and attitude but poorly translated into good complementary feeding behavior in the post-conflict development phase. Poverty, food insecurity and maternal ill health are the major factors that hinder care giver's efforts to adhere to recommended (good) complementary feeding practices. Maternal health, food security improvement, and poverty reduction should be prioritized if good complementary feeding is to be achieved among conflict-affected communities in the post-conflict development phase.

ACKNOWLEDG MENT
The authors are grateful to the Regional Universities Forum for Capacity Building in Agriculture (RUFORUM) for funding the study under grant numbers RU 2014 NG 15, RU 2014 NG 13, and RU 2014 GRG-098.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

E TH I C A L R E V I E W
The study was approved by the Gulu University Research Ethics Committee (GUREC/02/07/2016). Participation in the study was voluntary. Therefore, full consent of participants was sought and each participant signed a consent form before engaging in the study.
Participants were assured of confidentiality of information provided, and codes were used for personal information such as names and location to ensure anonymity.