UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Prevalence of iron deficiency and correlates of mild and moderate anaemia in infants six to eleven months… Daly, Zachary 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_february_daly_zachary.pdf [ 2MB ]
Metadata
JSON: 24-1.0167049.json
JSON-LD: 24-1.0167049-ld.json
RDF/XML (Pretty): 24-1.0167049-rdf.xml
RDF/JSON: 24-1.0167049-rdf.json
Turtle: 24-1.0167049-turtle.txt
N-Triples: 24-1.0167049-rdf-ntriples.txt
Original Record: 24-1.0167049-source.json
Full Text
24-1.0167049-fulltext.txt
Citation
24-1.0167049.ris

Full Text

  PREVALENCE OF IRON DEFICIENCY AND CORRELATES OF MILD AND MODERATE ANAEMIA IN INFANTS SIX TO ELEVEN MONTHS FROM MBALA DISTRICT, NORTHERN PROVINCE, ZAMBIA by ZACHARY DALY  B.Sc.(Agr.) , The University of Guelph, 2010  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Integrated Studies in Land and Food Systems) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  November 2014 © Zachary Daly, 2014ii  Abstract BACKGROUND: Childhood anaemia, defined as haemoglobin (Hb) < 110 g/L, is widespread in Zambia, affecting 55% of children under 5 years and 70% of children under 12 months. More broadly, it is a global concern that has been linked to a range of negative health outcomes, some irreversible. While the causes of anaemia are well understood, the specific etiology varies regionally, complicating efforts to create interventions. Unfortunately, there has been only limited research on the causes of anaemia in Zambia. OBJECTIVES: To determine the prevalence of anaemia, iron deficiency, and iron deficiency anaemia (IDA) in children 6-11 months in Mbala District, Northern Province, Zambia. Exploratory objectives examined factors associated with mild and moderate anaemia and haemoglobin concentrations to inform future interventions and research. METHODS: Analysis was performed on a convenience sample of 631 child-caregiver pairs that were recruited in Mbala District. Data on demographics, morbidity, sanitation and hygiene, and infant and young child feeding (IYCF) practices, anthropometry and the presence of malaria parasites was collected. Blood samples were obtained to measure haemoglobin, serum ferritin, serum transferrin receptor (STfR), and indicators of inflammation. RESULTS: 57% of the children were anaemic, 74% exhibited inflammation and the rate of iron deficiency varied between 13% when determined using inflammation adjusted serum ferritin cut-offs and 93% when using STfR > 8.3 mg/L, resulting in rates of IDA ranging from 8-22%. Using logistic regression it was found that increased age and achieving minimum dietary diversity were associated with reduced risk of being mildly or moderately anaemic, while increased malaria parasite loads, STfR concentrations, C-reactive protein concentrations, fever in the previous 2 weeks, and having been counselled on IYCF were associated with increased risk (p < 0.05). iii  CONCLUSION: Anaemia is widespread, representing a major health burden. While there is potentially widespread iron deficiency, inflammation makes it difficult to state what the true prevalence is, suggesting the need for further research. Interventions targeting dietary diversity, malaria and other infections, and iron status may be effective for lowering anaemia in the area. Future studies measuring further covariates and establishing causal pathways are required. iv  Preface   This thesis is based on research conducted by myself, the candidate, Zachary Daly, under the supervision of Dr. Judy McLean with guidance from my supervisory committee members Dr. Tim Green, and Dr. Larry Lynd.  The research presented in this thesis is part of a larger project investigating the effect of Micronutrient Powders on Zambian children 6-23 months in Mbala District, Northern Province, funded by Irish Aid and UNICEF and lead by the Zambian Ministry of Health. The study design, logical framework, and indicator matrix were completed by Agnes Aongola, Chief Nutrition Liaison Officer from the Ministry of Health, Dr. Judy McLean, Melanie Suter, Dr. Tim Green, and myself.  Data collection and entry was conducted by myself, Martina Northrup-Lyons, Diane Kim, and a team of Zambian enumerators. Agnes Aongola from the Ministry of Health, Dominique Brunet and Ruth Syandi from UNICEF-Zambia, Daniel Chiluba from the Provincial Medical Office and Bellington Kangaya from the District Medical Office provided key guidance in country. Freddie Mubanga, Eustina Mulenga Besa, and Jossy Phiri from the National Food and Nutrition Council performed community mobilization.  Justin Chileshe and a team from the Tropical Disease Research Centre (TDRC) were responsible for the collection of blood samples which were analyzed by the TDRC in Zambia and the Erhardt Laboratory in Germany. The analysis and writing of this thesis are my work. Sections of this thesis will be submitted for publication. Ethical approval was obtained from the Tropical Disease Research Centre Ethics Review committee in Zambia, approval number TRC/C4/05/2013, and the University of British Columbia’s Clinical Research Ethics Board, approval number H13-00261. The study was registered with ClinicalTrials.gov as NCT01878734. v  Table of Contents Abstract ........................................................................................................................................... ii Preface............................................................................................................................................ iv Table of Contents ............................................................................................................................ v List of Tables ............................................................................................................................... viii List of Figures ................................................................................................................................. x List of Abbreviations ..................................................................................................................... xi Acknowledgments......................................................................................................................... xii 1. Introduction ................................................................................................................................. 1 1.1. Objectives ............................................................................................................................ 4 1.2. Organization of the Thesis .................................................................................................. 6 2. Literature Review........................................................................................................................ 7 2.1. Anaemia and its Diagnosis .................................................................................................. 7 2.2. Anaemia and its Causes .................................................................................................... 10 2.2.1. Dietary Causes of Anaemia ....................................................................................... 10 2.2.2. Iron Deficiency Anaemia (IDA) ................................................................................ 12 2.2.3. Other Micronutrient Deficiencies and Anaemia ........................................................ 14 2.2.4. Non-Nutritional Causes of Anaemia ......................................................................... 14 2.2.5. Non-Causal Factors Associated with Anaemia ......................................................... 18 2.3. The Global Context of Anaemia: Prevalence and Consequences ..................................... 19 2.3.1. Iron Deficiency Anaemia (IDA) and Early Life........................................................ 21 2.4. The Zambian Context of Childhood Anaemia .................................................................. 22 2.4.1. Causes of Childhood Anaemia in Zambia ................................................................. 24 2.5. Measurement of Iron Deficiency ...................................................................................... 29 2.5.1. Bone Marrow Examination ....................................................................................... 30 vi  2.5.2. Serum Ferritin ............................................................................................................ 30 2.5.3. Serum Transferrin Receptor (STfR) .......................................................................... 33 2.5.4. Serum transferrin (STfR), Log Ferritin Index, and the Cook Equation .................... 36 2.6. Summary of Key Gaps in the Literature ........................................................................... 37 3. Methods..................................................................................................................................... 39 3.1. Sampling and Participants ................................................................................................. 39 3.2. Ethical Consideration ........................................................................................................ 42 3.3. Inclusion and Exclusion Criteria ....................................................................................... 42 3.4. Evaluation Tools ............................................................................................................... 43 3.4.1. The Questionnaire ..................................................................................................... 43 3.4.2. Anthropometric Data ................................................................................................. 45 3.4.3. Blood Collection ........................................................................................................ 46 3.5. Data Analysis and Statistics .............................................................................................. 48 4. Results ....................................................................................................................................... 51 4.1. Recruitment ....................................................................................................................... 51 4.2. Household and Participant Characteristics ........................................................................ 51 4.3. Morbidity and Health Characteristics ............................................................................... 52 4.4. Infant and Young Child Feeding (IYCF) Knowledge and Practices ................................ 53 4.5. Anthropometry .................................................................................................................. 54 4.6. Vitamin A Deficiency, Anaemia, Iron Deficiency, and Iron Deficiency Anaemia .......... 57 4.7. Bivariate Analysis of the Factors Associated with Haemoglobin Concentration ............. 59 4.8. Bivariate Analysis of the Factors Associated with Anaemia Status ................................. 63 4.9. Linear Modelling of Haemoglobin Concentrations .......................................................... 67 4.9.1. Assumption Checking of the Linear Model .............................................................. 68 4.10. Logistic Modelling of Anaemia Status ........................................................................... 70 vii  4.10.1. Assumption Checking of the Logistic Model .......................................................... 71 5. Discussion ................................................................................................................................. 73 5.1. Prevalence of Anaemia ..................................................................................................... 73 5.2. Prevalence of Iron Deficiency and Iron Deficiency Anaemia .......................................... 74 5.3. Participant and Household Characteristics ........................................................................ 76 5.4. Factors Associated with Haemoglobin Concentration and Anaemia Status ..................... 78 5.5. Modelling of Haemoglobin Concentration and Anaemia Status ...................................... 83 5.6. Limitations ........................................................................................................................ 85 5.7. Future Research ................................................................................................................. 87 6. Conclusion ................................................................................................................................ 89 Bibliography ................................................................................................................................. 91 Appendix A: List of Selected Zones, by Catchment Area .......................................................... 101 Appendix B: Consent Forms ....................................................................................................... 102 Appendix C: Questionnaire......................................................................................................... 106   viii  List of Tables  Table 3.1  Serum ferritin cut-offs used to diagnose iron deficiency in for each inflammation phase, as determined by C-reactive protein and alpha-1-acid glycoprotein concentrations ......... 48 Table 4.1 Household and participant characteristics .................................................................... 52 Table 4.2 Morbidity and health of children 6-11 months, in Mbala District ............................... 53 Table 4.3 Infant and young child feeding (IYCF) knowledge and practices of caregivers in Mbala District ............................................................................................................................... 54 Table 4.4 Stunting, wasting, underweight and mid-upper arm circumference (MUAC) of male and female children 6-11 months, in Mbala District .................................................................... 56 Table 4.5 Haemoglobin (Hb) and anaemia of children 6-11 months, in Mbala District ............. 57 Table 4.6 Iron deficiency in children 6-11 months, in Mbala District, using different definitions....................................................................................................................................................... 58 Table 4.7 Iron deficiency anaemia (IDA) in children 6-11 months, in Mbala District, using different definitions ....................................................................................................................... 59 Table 4.8 Bivariate comparisons of haemoglobin concentration against continuous variables in a sample of 6-11 month old infants in Mbala District ..................................................................... 60 Table 4.9 Bivariate comparisons of haemoglobin (Hb) concentration against categorical household and participant characteristics in a sample of 6-11 month old infants in Mbala District....................................................................................................................................................... 61 Table 4.10 Bivariate comparisons of haemoglobin (Hb) concentration against categorical morbidity and health characteristics in a sample of 6-11 month old infants in Mbala District .... 62 Table 4.11 Bivariate comparisons of haemoglobin (Hb) concentration against categorical anthropometric and biochemical characteristics in a sample of 6-11 month old infants in Mbala District........................................................................................................................................... 63 Table 4.12 Bivariate comparison of anaemia status against continuous participant, household, morbidity, anthropometric and biochemical characteristics in a sample of 6-11 month old infants in Mbala District ........................................................................................................................... 64 Table 4.13 Bivariate comparison of anaemia status against categorical participant, household, morbidity, anthropometric and biochemical characteristics in a sample of 6-11 month old infants in Mbala District ........................................................................................................................... 66 ix  Table 4.14 Linear Regression Model of Haemoglobin (Hb) Concentrations (g/L) of Children 6-11 Months in Mbala District, Zambia ........................................................................................... 67 Table 4.15 Logistic Regression Model of Anaemia Status of Children 6-11 Months in Mbala District, Zambia ............................................................................................................................ 71  x  List of Figures Figure 3.1 Map Showing Catchment Areas in Mbala District, Northern Province, Zambia, Which Were Selected for Sampling. ............................................................................................. 41 Figure 4.1 Standardized Predicted Haemoglobin Values versus Standardized Residuals of a Linear Regression Model of Haemoglobin (Hb) Concentrations of Children 6-11 Months in Mbala District, Zambia ................................................................................................................. 69 Figure 4.2 Normal P-P Plot of Regression Standardized Residuals of a Linear Regression Model of Haemoglobin (Hb) Concentrations of Children 6-11 Months in Mbala District, Zambia ....... 70  xi  List of Abbreviations ACD   Anaemia of chronic disease ACI   Anaemia of chronic inflammation AGP   Alpha-1-acid glycoproteini BIS   Body iron store CDC   Centers for Disease Control and Prevention CI   Confidence Interval CRP   C-reactive protein DHS   Demographic health survey EDTA   Ethylenediaminetetraacetic acid ELISA   Enzyme-linked immunosorbent assay GDP   Gross domestic product HAZ   Height-for-age z-score Hb   Haemoglobin IDA   Iron deficiency anaemia IL-1   Interleukin 1 IL-6   Interleukin 6 INACG  International Nutritional Anemia Consultative Group  INFγ   Interferon gamma ITN   Insecticide treated net IYCF   Infant and young child feeding MCDMCH  Ministry of Community Development Mother Child Health MCH   Mean corpuscular haemoglobin MCHC  Mean corpuscular haemoglobin concentration MCV   Mean corpuscular volume MNP   Micronutrient powder MoH   Ministry of Health MUAC  Mid-upper arm circumference RBP   Retinol binding protein SES   Socioeconomic status STfR   Serum transferrin receptor SUN   Scaling Up Nutrition TDRC   Tropical Disease Research Centre TIBC   Total iron binding capacity TNF   Tumour necrosis factor UN   United Nations WAZ   Weight-for-age z-score WHO   World Health Organization WHZ   Weight-for-height z-score YLD   Years lived with disability  xii  Acknowledgments First and foremost I would like to thank Dr. Judy McLean not only for taking me on as a student and acting as my supervisor, but for the many rewarding opportunities she has provided me. Her guidance, thoughts and advice have been instrumental in completing this thesis. I would also like to thank Dr. Tim Green for his ongoing support through the various stages of my degree and of the writing of this thesis, as well as Dr. Larry Lynd, for his valued assistance in my statistics generally and my modelling specifically. I am grateful for the assistance provided by the Faculty of Land and Food Systems, in the form of the David and Mary Macaree Fellowship as well as the funding provided by the Canadian Institutes of Health Research. Without this funding I would not have been able to participate in the program. I thank my lab mates Kyly, Kristina, Vashti, Phil, Crystal, Abeer, Amyna, and Allie, for making this experience so enjoyable. I also thank Barb, Lia, and Shelley for being patient and always being available. Many thanks go to the project participants: Agnes Aongola from the Zambian Ministry of Health, Dominque Brunet and Ruth Syandi from UNICEF-Zambia, Martina Northrup-Lyons, Diane Kim, and Melanie Suter. I also thank the many individuals from the TDRC, the Erhardt Laboratory and the NFNC, as well as the Provincial and District Medical Offices, for all of their support and hard work, as well as the funding from Irish Aid and UNICEF. Heartfelt thanks go to the team of enumerators for gathering the data, and most importantly I thank the caregivers involved in the project. Last, but by no means least, my deepest thanks go to my parents Jim and Kate for everything they have provided me in working towards and completing this program. Without your love and support none of this would have been possible.1  1. Introduction There are currently over one million malnourished children under the age of five in Zambia (1). Recognizing the high burden of malnutrition in the country, the Republic of Zambia joined the Scaling Up Nutrition (SUN) Movement of the United Nations in 2011, one of the first countries to do so (2). Parallel to this, Zambia has implemented a national programme focusing on early life nutrition, known as the First 1000 Most Critical Days Programme (1). This programme focuses on the “critical window” from conception, through birth and up to two years of age during which a failure to meet nutritional needs can lead to serious and irreversible consequences (1). The need for such programmes is borne out by national statistics: Zambia has the 13th highest under-5 mortality rate globally, a rate which increased from 163 deaths per 1000 children in 1990 to 170 in 2007 (3). Indicators of nutritional status are also troubling: the latest Demographic Health Survey (DHS) found that 45% of children under five exhibited stunting, or low height-for-age, 5% were wasted, with low weight-for-height, and 15% were underweight, with low weight-for-age (4). The rate of stunting in Zambia makes it one of only 18 nations in the world with a stunting rate over 45% (5). Furthermore, the average annual reduction in underweight from 1990 to 2008 was only 2%, reflecting “insufficient progress” (5), making it unlikely that Zambia will achieve Millennium Development Goal 1c, the reduction of underweight by 50% in children under 5 years (6).  However, one of the most troubling nutritional indicators, and the focus of the present thesis, is the prevalence of anaemia: as of the 2012 National Malaria Indicator Survey, 55% of children under five were anaemic, defined as haemoglobin (Hb) concentrations of less than 110 g/L (7). The prevalence is even higher in children under 12 months, at 70% (7). This puts the prevalence of anaemia in this population above the 40% cut-off the World Health Organization 2  (WHO) uses to designate a  “severe public health problem” (8). Furthermore, the most recent prevalence numbers represent an increase above the 2003 estimate of 49% (9). The consequences of anaemia, and particularly iron deficiency anaemia (IDA), are serious. This is particularly the case for children 6-23 months as during this time nutritional requirements are increased to sustain rapid growth (10). Iron deficiency in children is associated with irreversible, negative impacts on intellectual and physical development, as well as increased morbidity and mortality (8). Furthermore, when rates of anaemia, and thus iron deficiency anaemia, are high, there is likely a significant impact on GDP and economic growth, due to lost work capacity of the nation’s labour force (11). The reasons for the high rate of anaemia are likely multifactorial in Zambia as elsewhere; there are a number of potential causes of anaemia, including chronic infection, and deficiencies of micronutrients such as B12 or folate (8). Plasmodium, or malarial infection (12), as well as the effects of soil transmitted helminthes, such as hookworms, and Schistosoma infections can also cause anaemia (13). However, the main cause of anaemia worldwide is presumed to be iron deficiency, and it is generally assumed that within a given population, iron deficiency is the cause of approximately half of the cases of anaemia (8). One recent Zambian study found a 47% prevalence of iron deficiency anaemia in 6 month old infants (14).  This is of relevance to the current situation, as proper infant and young child feeding (IYCF) practices, particularly exclusive breastfeeding for the first six months, are generally considered to be protective against various nutritional deficiencies, including iron deficiency, during early life (15). Why then is anaemia in Zambia already so high in children under twelve months? Part of the answer may be IYCF practices themselves, while 94% of 6-8 months are breastfed in tandem with complementary foods, only 61% of children under six months are exclusively breastfed (4). 3  Broadly speaking, while there was progress in improving child nutrition during the 1990s, since 1999 this has stalled and as of 2003 the level of child malnutrition had “deteriorated”, partly due to the impact of a series of droughts on the food supply and maternal micronutrient status (16) and the impact of HIV/AIDS (9), which as of 2007 infected 14% of the population (17). Of note, 61% of the population are considered to be below the poverty line, and 42% are considered to be living in extreme poverty, rates which remain virtually unchanged from 2006 (18). Also of concern is the high prevalence of malarial infection, which was found in 15% of children under five during the most recent Malaria Indicator Survey (7), a prevalence which is considered mesoendemic according to WHO standards (19).  While there is limited data on the prevalence of other parasitic infections, a recent study in the Eastern Province found a 6% prevalence of Taenia solium, a type of tapeworm, a prevalence which is indicative of hyperendemicity (20). However, a 1994 study of children in Luapala Province found only a 2% prevalence of Ascaris lumbricoides (21). It is also estimated that nearly 2 million Zambians suffer from Bilharzia, or Schistosomiasis infection (22). Unfortunately, there is generally a lack of either up to date or nationally representative data on the prevalence on these and other neglected tropical diseases. More broadly, there has been little work done on the specific etiology of anaemia in Zambia, and even within a country one would expect reasons to be context and location specific; the causes of anaemia in a rural setting may very well differ from those in an urban one. Thus, the causes for the high prevalence of anaemia in the country, particularly in children under 12 months, remains an open question, one which must be answered to assist in the design of targeted interventions. This thesis will add to this literature by performing analysis, and looking at potential factors associated with anaemia that have not yet been analyzed in the Zambian 4  context. While it is not possible to establish causation, the results should prove useful in designing future studies, as well as to confirm previous findings. Most importantly though, this information will be useful in the design in future programmes and interventions aimed at improving nutrition during the first 1000 days and beyond. Efforts to understand anaemia in Zambia are complicated by the fact that there is a high burden of both acute and chronic disease (4). This makes it is very difficult to assess iron deficiency, as there is still a debate in the literature as to its assessment under such conditions (23). In fact, recent work has shown that depending on the assessment method chosen, one can arrive at very different estimates of the prevalence of iron deficiency within the same population sample (24). This is problematic for two key reasons. On the one hand an under-estimate of the prevalence of iron deficiency may result in poor targeting of any interventions seeking to reduce anaemia (24). However, the reverse, over-assessment of iron deficiency is also problematic given the concerns about giving iron to iron replete individuals in a malaria zone (25). There is also very little information on the prevalence of iron deficiency in the country. Taken together, this forms the basis for other questions which need to be asked: what is the rate of iron deficiency, and thus iron deficiency anaemia in Zambia? Further, do the different definitions of iron deficiency find similar or dissimilar results in the Zambian context? This last question is also relevant beyond the Zambian context, as it will inform the use of these measures in other settings, if the findings of previous studies showing disagreement are replicated.  1.1. Objectives Fortunately, there are a number of interventions that can be used when tackling childhood anaemia in a population, but one that has been found to be particularly effective in a variety of settings has been Micronutrient Powders, or MNP (26–28). With this in mind, the Zambian 5  Ministry of Health, in partnership with Irish Aid, UNICEF-Zambia and the University of British Columbia decided to pilot a project in Mbala District, Northern Province, which made use of MNP, de-worming tablets, insecticide-treated nets (ITN), and training for caregivers on infant and young child feeding (IYCF).  While the project included a two-armed study protocol, the present thesis will focus on analysis of the baseline data, treating it as a cohort study, and ignoring whether a given participant was part of the control or treatment arms. Furthermore, given that children with severe anaemia were excluded from the larger study for safety and programmatic reasons, I can only comment on mild and moderate anaemia, as there are no children with severe anaemia in the dataset. Thus, the objectives of this research are to: i. Determine the prevalence of mild and moderate anaemia in children 6-11 months in Mbala District, Northern Zambia; ii. Determine the prevalence of iron deficiency, and thus iron deficiency anaemia, in children 6-11 months in Mbala District, Northern Province using different methods of biochemical assessment; iii. Determine the factors associated with mild and moderate anaemia, in children 6-11 months in Mbala District, Northern Province, to better understand the potential causes of anaemia, focusing on biochemical indicators, household characteristics and IYCF practices; iv. Determine factors associated with haemoglobin concentration, in children 6-11 months in Mbala District, Northern Province, to better understand the potential 6  causes of anaemia, focusing on biochemical indicators, household characteristics and IYCF practices 1.2. Organization of the Thesis The remainder of this thesis is divided into four chapters: a literature review which will discuss the relevant literature not only on anaemia in the Zambian context, but also the broader literature about the causes of anaemia, as well as the current thinking on the biochemical assessment of iron deficiency. This will be followed by a methods section, which will explain study design and sampling, as well as data collection and statistical analysis. A chapter on the results will discuss not only the calculations of the prevalence of iron deficiency, but also the different factors associated with mild and moderate anaemia. Next, a discussion chapter will discuss the implications of the findings, as well as some of the limitations. Finally, a conclusion chapter summarizing the key findings, while also discussing future avenues for research.   7  2. Literature Review  I will start by looking at the broader literature concerning anaemia: its diagnosis, and causes, along with the global context of the condition. I will then turn to available literature on childhood anaemia in Zambia, its prevalence, and the current understanding of its causes. I will conclude with a section on the diagnosis of iron deficiency, with an emphasis on the biochemical assessment of iron deficiency in children under conditions of a high burden of disease, as seen in Zambia. Finally, I review of the key gaps in the literature which this thesis will endeavor to address. 2.1. Anaemia and its Diagnosis Anaemia is defined as “a condition in which the number of red blood cells (and consequently their oxygen carrying capacity) is insufficient to meet the body’s physiological needs” (29). The biochemical assessment of anaemia makes use of haemoglobin concentrations in the blood, and thus a more technical definition of anaemia is a concentration of haemoglobin below the 2.5th percentile in a healthy population (30). Haemoglobin is responsible for the transport of oxygen from the lungs to the rest of the body, and is found in red blood cells. It is also accounts for 70% of the body’s stores of iron, which is important for haemoglobin structurally (31). When measuring haemoglobin, venous blood  stored in ethylenediaminetetraacetic acid (EDTA) is preferred, although it is also possible to use capillary blood obtained by pricking a finger, heel or ear; however when using capillary blood, the sample may by contaminated with interstitial fluid, which may result in an erroneously low haemoglobin concentration (31). Alternatively, it is possible to assess haemoglobin using blood from dried blood spots (31).  8  Once collected there are two main methods to characterize haemoglobin concentrations. One is the cyanomethemoglobin method, which is considered very reliable, with a very low coefficient of variation (31). However, a more field friendly method makes use of the HemoCue, which is a portable and battery operated photometer (31). When assessed properly, venous samples analyzed using a HemoCue are comparable to the cyanomethemoglobin method (31). The WHO, Centers for Disease Control (CDC), UNICEF, and the International Nutritional Anemia Consultative Group (INACG) have a set of recommended cut-offs for the assessment of anaemia using haemoglobin concentrations (32). The cut-offs are broken into different age categories and above the age of 14 there are different cut-offs for men, non-pregnant women and pregnant women. The cut-off for anaemia in children 6-59 months is Hb < 110 g/L and is not sex specific (29). A concentration of Hb < 70 g/L is defined as severe anaemia and a concentration of Hb < 100 g/L is defined as moderate anaemia (28). It should be noted that these cut-offs are not without controversy, particularly those for children, as they were determined by extrapolating from older groups (31). As such there are a number of alternate cut-offs, such as those by Domellof et al. which suggest a cut-off of Hb < 100g/L for children that are 9 months old (33), Emond et al. which suggest a cut-off of Hb < 97g/L in British infants that are 8 months old, based upon capillary haemoglobin levels, (34) and Thorosdottir who argues for Hb <105g/L in Icelandic infants 12 months of age (35). However, none of these alternate cut-offs have been adopted internationally, and so will not be discussed further. Furthermore, there are a number of correction factors which can be applied to the WHO cut-offs for different situations depending on the race, altitude, and smoking status of the individual being assessed (30). The use of race based cut-offs was based on the finding that even when iron replete, individuals of African descent seemed to have haemoglobin concentrations 5-9  10 g/L lower than individuals of Caucasian descent, even when controlling for age and income (36). Notably, however, the traditional cut-offs do seem to be appropriate for other ethnic groups, such as those from Indonesia (37).  From these findings the WHO has in the past recommended that the cut-offs should be lowered by 10 g/L, regardless of age, when assessing anaemia in individuals of African descent (23). However, more recent guidelines from the WHO state the opposite, as “the data is still scarce,” and call for the use of the original, unadjusted, cut-offs (29). In line with this, it is very common for studies on African populations to use unadjusted cut-offs, including studies and reports set in Zambia (9,21,38,39). It is also important to note that the original studies indicating a difference in iron replete haemoglobin concentrations were based upon data from African Americans living in the United States, rather than in individuals of African descent still living in Africa, where one would expect the causes and etiology of anaemia to be different (36). Furthermore, even when race based adjustments are recommended, there is disagreement on how large they should be; the Institute of Medicine has in the past suggested that when diagnosing anaemia in American children and adults of African descent the cut-offs be lowered by 0.3 g/L and 0.8 g/L respectively, notably lower than the earlier, now defunct, recommendation from the WHO (40). Another important consideration when setting a cut-off for diagnosing anaemia is the altitude at which a person resides. This is because as elevation increases so does hematocrit and haemoglobin concentrations in response to the lower oxygen concentration found in the air (31). There are currently two different sets of correction factors, one put out by the WHO, and one by the INACG, which vary slightly (31). However, both recommend the use of a correction factor at elevations above 1000 metres. Notably, the correction factors for elevations past 3000 metres 10  exist, but were extrapolated from values determined at lower elevations. However, like the race based adjustments previously mentioned, use of these correction factors is far from universal, including in Zambia, which does not make use of them in national surveys (7,39,41,42). 2.2. Anaemia and its Causes There are many different causes of anaemia, and the precise etiology of the condition varies widely from place to place, which can make it difficult to generalize findings from one location to another. A useful framework for understanding these different causes comes from Thurnham and Northrop-Clewes which divides the causes of anaemia into three broad clusters: food, disease leading to inflammation, and blood loss (43). Examples of the first would include iron deficiency anaemia (IDA), along with other nutritional anaemias, and examples of the final include hookworms, schistosomiasis, and malaria (43). Disease leading to inflammation refers to anaemia of chronic disease (ACD), sometimes referred to as anaemia of chronic inflammation (ACI) (43). In addition to these three broad exogenous causes, anaemias can also be induced by endogenous factors, notably genetic disorders such as sickle cell condition or thalassemia, which impair the ability of red blood cells to adequately fulfill their physiological role, either due to an ability to properly transport oxygen or due to inadequate production or excessive destruction relative to needs (43). 2.2.1. Dietary Causes of Anaemia Key in the understanding of anaemia are the so called nutritional anaemias, those which are characterized by an inadequate dietary intake or absorption of one of more micronutrients. Nutritional anaemias are often given the most attention, particularly in the development of interventions designed to reduce rates of anaemia in the developing world; however some have 11  questioned this focus, suggesting it may be partially behind the inability to adequately reduce global anaemia rates (8).  In addition to assessing individual micronutrients, which may be related to anaemia, it is also possible to assess the overall quality of the diet, using standardized indicators of IYCF practices. Three of the key indicators are minimum meal frequency, which measures whether children are receiving enough meals per day, minimum dietary diversity, which whether children are consuming food from a minimum number of food groups, and minimum acceptable diet, which assesses a combination of both (44). Dietary diversity has been found to be a good indicator of micronutrient density of a diet and its overall quality in a number of settings (45,46). The WHO also recommends that up to six months of age a child should be exclusively breastfed, with breastfeeding continuing up to two years of age; however, at six months breast milk is no longer sufficient to meet all of the needs of a growing infant and so complementary foods must be introduced (15). In fact, during the period from 9-11 months of age breast milk is only able to provide 3% of the iron needs of the infant, the rest must come from complementary foods (15). Because of this, it is recommended that children consume iron-rich foods, such as meat or fish every day (15), which is the basis for another key IYCF indicator: consumption of iron-rich foods in the previous 24 hours (44). Unfortunately breastfeeding practices in Zambia are not ideal, with only 61% of children under six months exclusively breastfed, with a median duration of exclusive breastfeeding only 3.1 months (4). The median length of continued feeding is only 20.3 months, in line with the finding that only 45% of children 18-23 months are still breast fed (4).  Furthermore, the Zambian diet is monotonous and largely based on maize as a staple, even within urban areas, to the point that fluctuations in the price of maize have been associated with changes in 12  anthropometry at a population level (16). It has also been suggested that increases in the price of maize prevent households from purchasing fortified foods, such as vitamin A fortified sugar (16). The overreliance on maize also makes it difficult for individuals, including children, to consume protein in adequate quantities and of sufficient quality (1).  This is borne out in the indicators of dietary quality: while 94% of children 6-8 months have started complementary feeding, only 37% of children 6-23 months achieve minimum dietary diversity, and only 49% the minimum meal frequency (44). When combined, only 25% of children 6-23 months are achieving the minimum acceptable diet, as defined by the WHO (44). Furthermore, while 65% of children 6-35 months consumed iron-rich foods in the previous day, this figure dropped to 43% when looking at children 6-8 months (44). Taken as a whole, this suggests that it would be very difficult for many infants in Zambia to meet their nutrient needs, particularly when it comes to micronutrients, such as iron which are key in the etiology of anaemia.  2.2.2. Iron Deficiency Anaemia (IDA) Iron deficiency resulting in IDA is believed to be the most common form of anaemia worldwide, with many authors stating that within a population approximately 50% of the given anaemia can be attributed to iron deficiency (47,48). Notably, this assertion has been challenged on the basis that the initial calculation was based upon a healthy North American population that is not comparable to others; however, this figure appears in UNICEF and WHO documentation (23,49), along with the research papers which cite these documents (48). A recent meta-analysis of the global anaemia burden concluded that, globally, IDA was the leading cause of anaemia in both men and women (50). However, the specific share of IDA in the overall etiology of anaemia varied widely by region, from as little as 3% in high-income North America to 65% in Central 13  Asia (50). Despite this even the WHO cautions that: “…the role of factors other than iron deficiency in the development of anaemia has been underestimated by public health officials, because for a long time anaemia has been confused with iron deficiency anaemia…” (8). It has also been suggested that the prevalence of iron deficiency in a population can be assumed to be 2.5 times whatever the prevalence of anaemia is (23). This implies that when the prevalence of anaemia surpasses 40%, one can assume that nearly all members of the population are iron deficient (23).  Key to understanding the relationship between iron deficiency and IDA is the fact that IDA is merely the end stage of iron deficiency: one can be iron deficient, but not yet anemic, just as someone could be anemic, but not suffering from any iron deficiency. Generally speaking the progression of iron deficiency leading to IDA is broken into three distinct phases (31). The first stage is iron depletion, in which the stores of iron within the liver and elsewhere in the body are drawn down as a result of intake falling short of needs (31). However, during this phase there are minimal or no changes to haemoglobin or transport iron concentrations, and as such there are few physiological or functional consequences (31). However, indicators of storage iron will be affected (31). The second stage, iron-deficient erythropoiesis, is sometimes referred to as iron deficiency without anaemia or early functional iron deficiency (31). During this phase the iron stores have become depleted and the supply of iron for red blood cell production begins to decline, resulting in a rise in serum transferrin concentrations (31). The final stage, IDA, sometimes referred to as established functional iron deficiency, occurs when the concentration of haemoglobin in the blood falls in response to inadequate iron supplies (31). Some authors also refer to another stage in the progression of IDA, proceeding iron depletion, known as negative iron balance (51). IDA is characterized by microcytic and hypochromic red blood cells with 14  lowered mean cell volume (MCV), lowered mean cell haemoglobin (MCH) and lowered mean cell haemoglobin concentration (MCHC) (31). 2.2.3. Other Micronutrient Deficiencies and Anaemia Other than iron deficiency, there are several other key nutritional deficiencies that can lead to anaemia. Two of the more important are deficiencies in folate and B12, both of which result in a macrocytic anaemia known as megaloblastic anaemia that is characterized by elevated MCV, elevated MCH, but normal MCHC (31). Deficiencies in vitamins A, B6, and riboflavin, as well as deficiencies in copper can also lead to anaemia (31). In 2003, it was found that 54% of Zambian children 6-59 months were vitamin A deficient, although this was an improvement over a previous estimate from 1997 of 66% (9). In terms of intake and supplementation, in 2007 it was estimated that 58% of children 6-8 months and 82% of children 9-11 months consumed foods rich in vitamin A in the previous 24 hours and 60% of children 6-59 months had received a vitamin A supplement in the previous six months, although rates of supplementation were lowest in Northern Province (4). Unfortunately there is little information available on intakes or biochemical indicators of the other micronutrients. 2.2.4. Non-Nutritional Causes of Anaemia A major cause of anaemia, particularly in the developing world, is anaemia of chronic disease (ACD), sometimes referred to as anaemia of chronic inflammation (ACI) now that it is better understood (43). It is characterized by normochromic and normocytic red blood cells, with normal values for MCV, MCH, and MCHC (31). During ACI, the chemical hepcidin increases in concentration, in response to increased levels of Interleukin-6 (52). Hepcidin then binds to the surface of RE macrophages, which are responsible for destroying old or damaged red blood cells as they pass through the spleen. The hepcidin degrades ferroportin, a transport protein for iron on 15  the surface of the macrophages (52). This causes iron within the macrophages to become trapped, effectively preventing the body from accessing the iron obtained from senescing red blood cells (52). Hepcidin also reduces the daily uptake of iron from the duodenum (52). If this process continues, as is the case in chronic inflammation, it will eventually result in IDA (52). This is potentially relevant to Zambia given the high burden of disease in children in the country: 7% of children under five showed symptoms of acute respiratory illness, 25% had a fever, and 18% had diarrhea during the previous two weeks, according to the most recent DHS (4). There are a number of parasites which can result in a loss of red blood cells through bleeding or other mechanisms, and thus anaemia. Key among these are malaria, Schistosomiasis, hookworms, Trichuriasis, and Ascaris. Malaria parasitizes red blood cells, destroying them, a process which is exacerbated by the still poorly understood removal of healthy red blood cells from the body by the immune system during infection (53). In Zambia, 10% of children under 12 months test positive for malaria parasites, in Northern Province it was found that 24% of all children under 5 years tested positive, which gives it the third highest prevalence of any province (7). While there is a 27% parasitism rate in children belonging to the lowest wealth quintile, this drops to 2% in the highest quintile (7). Overall, the trend has been a decrease in malaria morbidity and mortality; however, there was a resurgence in 2009-2010, associated with delays in funding for malaria control programs, particularly in the north-east of the country (54). The major methods for reducing the spread of malaria within the country are indoor residual spraying, and long-lasting insecticide-treated nets, of which 24.6 million were distributed from 2006-2011 (55). Another important factor to consider when assessing anaemia is the prevalence of Schistosomiasis, or bilharzia, in the form of S. haematobium and S. mansoni. While there is not 16  an up-to-date report on the prevalence of this disease, it is currently estimated to effect 2 million individuals (22). Mbala district is assumed to have a high prevalence of the disease, based on records from health facilities, with 1037 reported cases from 1999 to 2005 (56). Within Mbala District, the disease seems to affect the most individuals in the Munyezi, Mwamba, Chozi, Isofu, and Chisanza areas (56). A 2011 study in Zambezi District, North-Western Province, found a 30% prevalence of S. haematobium and a 15% prevalence of S. mansoni in children 7-19 years of age (57). There is little data on other parasites in Zambia, particularly at a national level; however, a 6% prevalence of Taenia solium, a type of tapeworm, was found the Eastern Province, indicative of hyperendemicity (20). A study of children in Luapala Province found only a 2% prevalence of Ascaris lumbricoides, although given that the study dates from 1994 it is difficult to know how this may relate to the present situation in Zambia (21), especially given that a more recent study found no evidence of the parasite in Zambezi District in 2011 (57). The study in Zambezi District also found that 43% of children had hookworm, and that there was a major issue of co-infection with 44% of the children in the study testing positive to multiple parasitic infection, including 9% who had malaria, hookworm, and S. haematobium (57). There is a de-worming program in place in Zambia, and according to the previous DHS, 60% of children 6-59 months had been given de-worming medication in the previous six months; Northern Province had the lowest rate at only 46% (4).  Within the category of endogenous causes of anaemia the most important conditions are haemoglobinopathies, such as sickle cell condition, and thalassemia (43). Unfortunately there is little nationally representative data on the prevalence of these conditions in Zambia, so it is 17  difficult to comment on their importance in the etiology of anaemia in the country. However, it is worth nothing that several of these conditions have been documented.  An analysis of 10,000 samples of Zambian blood conducted in 1969 found Haemoglobin S, resulting in 187 cases of sickle-cell anaemia, one case of Haemoblobin J, and four cases of a novel haemoglobin variant termed Haemoglobin Zambia (58). Furthermore, there was evidence of α-thalassaemia in the form of Haemoglobin Bart’s and Haemoglobin H disease (58). There was no evidence of β-thalassaemia (58).  Further evidence of α-thalassaemia was found in a 1989 study that analyzed 109 Zambian newborns from three different hospitals, representing a total of six different ethnic groups (59). Of the 109 infants, 38.5% were heterozygous for α-thalassaemia and 7.3% were homozygous for it (59). The mean haemoglobin concentration of non-carriers was 16.2 ± 2.1 g/L, as opposed to 14.93 ± 2.0 g/L for heterozygous individuals (p = 0.05) (59). The mean haemoglobin concentration of homozygous α-thalassaemia infants was 13.43 ± 2.0 g/L (p < 0.05) (59). It was also found that there was no difference in prevalence of the gene for α-thalassaemia by ethnic group, and that all instances of the α-thalassaemia trait consisted of –α3.7, with no instances of – α4.2 (59). Finally, a 1983 study in Nchelenge District, Luapula Province, found the sickle cell trait in 70 of 424 pregnant women of low socioeconomic status screened who were at antenatal clinics (60). Interestingly, the study found no relationship between the haemoglobin concentrations of the women and whether or not they had the Haemoglobin S trait (60). 18  2.2.5. Non-Causal Factors Associated with Anaemia  A number of studies have found associations with so-called “distal variables” and the risk of having anaemia, although findings are very inconsistent from place to place. The household wealth index was found to be associated with anaemia in Nigeria, but this was not the case for the maternal level of education (61). This contrasts with studies from Tanzania, one of which found a statistically significant relationship between maternal education and anaemia (62), and another that found no association between household wealth and anaemia (63). Along a similar vein, a study from Indonesia found that maternal knowledge of anaemia was associated with reduced risk of childhood anaemia in rural settings (48). Other variables which have been associated with childhood anaemia are a lack of household toilet facilities (63) and maternal haemoglobin concentrations (64,65). While there appears to be only a weak association at most between the macronutrient composition of milk and maternal diet or body composition (66), severely anaemic mothers have been found to have lower quantities of iron in their breast milk, which could have consequences for infant iron status, and thus the risk of anaemia (67). Furthermore, newborns of mothers with low iron stores may themselves be born iron deficient, and these children have been found to have significantly lower iron stores at nine months compared to children who were born with adequate iron stores (68).  According to the DHS, 10% of Zambian women, and 13% of those in Northern Province have received no formal education (4). Furthermore, only 24% of households have access to an improved latrine or toilet facility, and numbers are even lower in rural settings (4), where, according to the 2012 Malaria Indicator Survey, 34% of households had no toilet facility at all, relying upon bushes or fields (7). The prevalence of anaemia among non-pregnant women of child bearing age has been placed at 29%, although this figure dates from 2003 (9). 19  2.3. The Global Context of Anaemia: Prevalence and Consequences The most recent global estimates of the prevalence of anaemia by the WHO are based on data from 1993-2005 and puts the global prevalence at 25%, which corresponds to over 1.6 billion people (8). When looking at preschool children (0 – 59 months) the prevalence climbs to 47%, the highest of any group, representing 293 million individuals. In comparison, 42%, 30%, and 13% of pregnant women, non-pregnant women, and men, respectively, are anaemic (8). Within Africa, the prevalence of anaemia among preschool aged children was estimated to be 65%, representing 93.2 million children (8). This is the highest prevalence rate of any UN region, although second to Asia in terms of absolute numbers (8). The WHO has a set of guidelines for assessing the severity of the public health problem posed by anaemia at a population level, which help to put these figures into context; a prevalence over 40% is considered to be of severe public health significance (23).  It is difficult to compare global prevalence across studies or time, due to changing methodologies and cut-offs (8). For instance, a 1985 study estimated the global prevalence of anaemia to be 30%; however it excluded all data from China. According to the WHO, it is not even possible to compare prevalence data from the 2008 report to estimates from a 1992 publication, itself also from the WHO, due to differences in methodology (8). Despite this, it is worthwhile to look at some more recent estimates, keeping in mind that they cannot be directly compared to the WHO estimate published in 2008. A meta-analysis from 2013 estimated that the global prevalence of anaemia among children under 5 had fallen from 47% to 43% between 1995 to 2011, although the confidence intervals for these estimates overlapped (47). Eastern Africa, the region containing Zambia, saw a fall from 74% to 55% (47). 20  The most recent global estimate, published in 2014, found that the prevalence of anaemia had dropped to 33% in 2010, from 40% in 1990 (50). As expected, when looking at different age groups the highest prevalence of anaemia was found in the post neonatal period, followed by those from 1 to 4 years of age (50). However, the paper used a non-standard anaemia cut-off of Hb <120 g/L for children under 5 years, which makes their results difficult to compare to other studies (69). It has been further pointed out that, while of value, these large “model-based” estimates of anaemia do “not replace the need for field epidemiology” (69). The concept of years lived with disability (YLD) can help put the global burden of anaemia in perspective. YLD are calculated by multiplying the prevalence of a given condition by its disability weight, which is an empirical measure of the impact of a condition on a given individual (50). These weights are derived from large surveys of affected individuals (50). Based upon global prevalence levels, it has been estimated that anaemia was responsible for 68.4 million YLD in 2010 (50). This represented 9% of the total YLD estimate for all conditions for that year (50). While this is in increase in absolute terms over the 1990 estimate of 65.5 million YLD, it represents an improvement in relative terms, as in 1990 anaemia was responsible for 11% of global YLD (50). Also of note was the decline in the overall disability weight of anaemia during this period. This fall was attributed in particular to reductions in severe anaemia among women, which has a higher disability weight than moderate or mild anaemia; however, there was not a commensurate decrease in the severity of anaemia in men (50). These disability weights at the individual level are the result of numerous consequences of anaemia including loss of appetite, fatigue, due to low levels of oxygen in the blood, an increased risk of maternal mortality during birth due to blood loss, poor regulation of body temperature, gastrointestinal disturbances and increased risk of infection by H. pylori (8,51). 21  2.3.1. Iron Deficiency Anaemia (IDA) and Early Life Anaemia resulting from iron deficiency, so called IDA, has repercussions above and beyond that of anaemia alone, particularly if IDA occurs during the first two years of life (10). This is because of the critical role of iron in the growth of the developing infant. In fact iron needs are at their highest during pregnancy and the age of 6-11 months and it is during the second year of life that the risk of deficiency is greatest (23).  These consequences are mostly centred around impaired motor and cognitive development (23).  Most troubling is that if not caught early and treated, the negative effects of iron deficiency may be irreversible, and in fact do not always respond to iron supplementation (70). In one landmark study looking at the effects IDA during infancy on Costa Rican children at 5 years of age, it was found that the children who had had IDA during infancy had significantly lower scores on Woodstock-Johnson Psycho-Educational Battery, as well as lower scores for the Beery Developmental Test of Visual-Motor Integration (71). Effects were most pronounced for tasks that made use of non-verbal skills, such as visual-motor integration (71). These findings remained even when the researchers controlled for the fact that many of the children with IDA came from lower socio-economic statuses (71). When these children were followed up at 19 years of age, it was found that those with iron deficiency during infancy had not attained equal cognitive scores as infants who had had adequate iron (72).  Other studies have found similar findings, as well as impaired social-emotional and neuro-physiological development (70). The precise consequences seem to vary depending on precisely when the deficiency occurred, reflecting the fact that different parts of the brain are developing at different points in time during infancy (70). A meta-analysis of studies looking at the long term effects IDA and cognitive development found that for every 10 g/L decrease in 22  haemoglobin, one could expect a 1.73 point drop in IQ (73). Of course, it is important to keep in mind that despite all of this evidence for the effect of iron deficiency on the developing brain, the precise cause is still uncertain. Potential explanations from rat models include effects on the hippocampus, the myelination of axons, and on various neurotransmitters (70). 2.4. The Zambian Context of Childhood Anaemia This section will relate the previous sections of the literature review to the specific situation in Zambia. The most recent nationally representative estimate of anaemia comes from the 2012 Malaria Indicator Survey, which found that 55% of children under 5 were anaemic (7). This rate was higher in children under 12 months, of which 70% were anaemic, the highest of any age cohort in the survey (7). It also reported that 7% of children were severely anaemia; however, this was based upon a non-standard cut-off of haemoglobin < 80 g/L, which goes against the current WHO standards (23), and is thus to be viewed as an over-estimate. This also makes it difficult to compare to other regions. Rural children had higher rates of anaemia, 59%, than urban children, of which 46% were anaemic (7). Furthermore, while the rate of anaemia decreased with increasing wealth, it was still 48% in the highest wealth quintile (7). The province with the highest prevalence was Eastern Province, and it was lowest in North-Western (7).  Northern province, the location of the MNP pilot, reported a 59% prevalence of anaemia in children under 5 (7). Overall, boys had a slightly higher prevalence of anaemia than girls (57% versus 53%), although there was no test for significance (7). The previous Malaria Indicator Surveys are a good source of information for understanding the trends in anaemia in the country, particularly because they use similar methodologies, allowing for comparisons across the years. The 2010 rate of anaemia in children 23  under 5 was 61%, and 69% in children under 12 months (39). The 2008 the overall rate of anaemia was 49%, with 62% in children under 12 months (42). Unfortunately, the 2006 survey did not report the overall prevalence of anaemia, but did report that 14% of children were severely anaemic, but again this was using the non-standard cut-off (41). There are older estimates of anaemia which were done as part of other surveys. A 2003 survey to evaluate the national vitamin A supplementation program found that 53% of children under 5 were anaemic (95% CI: 46.7, 59.3) (9). Of note was a seasonal pattern to the findings: in July 59% of children were anaemic, compared to 47% in November (9). It was found that 2% of children were severely anaemic, and, unlike the Malaria Indicator Surveys this was determined via the standard WHO recommended cut-off (9).  Finally, there is a 1998 survey focused on the prevalence and causes of anaemia in Zambia in different age groups (74). The study found that 65% (95% CI: 62, 67) of children under 5 were anaemic and 15% were severely anaemic (74). Unfortunately, this was the first assessment of the prevalence of anaemia that was undertaken in the country (74) and so it is not possible to look back further. Thus the overall trend in childhood anaemia rates in Zambia is not clear. While the most recent estimate is an improvement from 1998 estimates, it is worse than 2008. This suggests that there is still much work to be done in the country. Perhaps most troubling is the fact that the prevalence of anaemia has recently increased from its low in 2008. The next section will turn to the question of the causes of anaemia in the Zambian context, as is understood from the existing literature. 24  2.4.1. Causes of Childhood Anaemia in Zambia Unfortunately there is not a large body of literature on the causes of anaemia specific to the Zambian context. This is further complicated in that much of the literature that does exist is not highly generalizable, and so it is difficult to draw conclusions from individual studies to other areas of the country. Another issue is the lack of studies of key variables, such as B12 or folate deficiency. Finally, even where there are large or representative samples, the study design is such that causation can rarely be stated, just correlation.  One of the most comprehensive studies on anaemia in Zambia was nested into the 2003 evaluation of the national vitamin A supplementation program. The study evaluated 390 households from 30 different randomly selected clusters from the 2001 Demographic and Health Survey; within each cluster 13 children were randomly selected (9). They measured a number of variables which they hypothesized to be associated with anaemia. These were: concentration of C-reactive protein (CRP), malaria parasitism, vitamin A deficiency, reported fever, cough and diarrhea, de-worming, vitamin A supplementation, living  in a household with vitamin A fortified sugar, demographics and the educational attainment of the mother (9).  There was no significant relationship between anaemia and reported cough, living in a household with fortified sugar, vitamin A supplementation, household demographics or the education of the mother (9). However they did find that children who had elevated CRP concentrations, defined as CRP > 5mg/L, as well as those with vitamin A deficiency had higher rates of anaemia (9). They also found that children with malaria parasites and those whose caregivers reported fever or diarrhea had a higher prevalence of anaemia than those who did not (9). Finally they noted children who had received de-worming tablets had lower rates of anaemia than those children who had not received them (9). 25  A 1998 study into the prevalence and causes of anaemia in Zambia by the National Food and Nutrition Commission also provides insight, even if it is somewhat dated. The study made use of 60 clusters, and within each cluster 25 households were randomly selected, with a total of 1427 children screened (74). They found that caregiver reported malaria, malaria parasitism, and living in a rural setting were all associated with an increased risk of anaemia (74). In addition to these large, nationally representative studies, there are a number of smaller studies that shed light on the causes of anaemia in Zambia. A 2011 study looking at insecticide treated nets (ITNs) to lower the rate of malaria infection  in Luangwa District supports some of the older findings suggesting a link between malaria and anaemia in Zambia (75). A randomly selected sample of 1190 children under the age of 5 was assessed and the authors reported that there was a significant (p < 0.001) relationship between malaria parasitism and severe anaemia, despite the fact that the district had achieved over 80% coverage with ITNs (75). There was no relationship found in either bivariate or logistic regression analyses between severe anaemia and whether a household owned an ITN, or whether a child slept under an ITN (75). However, when interpreting the findings it is important to keep in mind that they used a non-standard definition of severe anaemia, haemoglobin < 80 g/L (75). Finally, as previously discussed, Zambian infants who are carriers for the α-thalassaemia trait have been found to have lower mean haemoglobin concentrations (59). The literature specific to iron deficiency in Zambian children is limited, and there is a lack of nationally representative studies. One study from 2007 sampled 38 children with malaria in four different clinics and hospitals, and assessed their haemoglobin concentration, and whether or not they were iron deficient (76). Of the 38 children, 10 (27%) were found to have no anaemia or iron deficiency, 2 (5%) were found to have iron deficiency but no anaemia, 13 (34%) were 26  found to have anaemia but not iron deficiency, and 13 (34%) were found to have concurrent anaemia and iron deficiency (76). Thus, approximately of a third of the children presented with iron deficiency anaemia. However, it is difficult to draw broader conclusions from this study for a number of reasons: it was a non-random sample, non-standard cut-offs were used to diagnose anaemia, the assessment of iron status did not use any biochemical indicators of iron, and it cannot be generalized to children without malaria (76). One way to assess whether anaemia is the result of iron deficiency is through the use of iron supplementation of anaemic children; one would expect to see an improvement in the concentration of haemoglobin if iron deficiency was the cause, and a failure to see an improvement would imply that another cause was responsible. One intervention in the Nangweshi Refugee Camp, one of several in Zambia, sought to reduce rates of anaemia and vitamin A deficiency through the replacement of traditional maize meal with locally fortified maize meal (77).  At baseline the rate of anaemia in children under 5 was 48% (95 CI: 39.7-55.9), one year later, post-intervention it had reduced to 24% (95% CI: 17.3-34.4), with a p-value < 0.001 (77).  Based upon this, as well as change in vitamin A deficiency seen in adolescent children, it was suggested that the maize meal had improved the micronutrient status of the children, implying a nutritional cause of the anaemia (77). However, even if these results are suggestive, there was no control group (77). As such, it is impossible to definitively state whether the maize meal was responsible for the change in anaemia prevalence, given that other factors could have changed in the camp during the year of the intervention (77). A better designed study was undertaken in the Macha area of southern Zambia (78). A total of 232 children with anaemia aged 18-120 months were enrolled and then randomized to 27  one of two treatment arms for three months: a control group which received a placebo syrup and an intervention group which received daily liquid ferrous sulphate (78). At the end of the three months haemoglobin was again measured, to see whether there had been a change in concentrations (78). While both groups saw an increase in haemoglobin, 21 g/L in the iron group and 17 g/L in the control group, the difference was not statistically significant (78). This lead the authors to conclude that within this particular population iron was not the leading cause of anaemia; rather another condition, which they hypothesized to be malaria, was (78). However, these results differ somewhat from an intervention using micronutrient fortified porridges (79). Six month old children in the capital, Lusaka, were split into two treatments arms: a rich porridge that had 19 micronutrients added to it (n = 373) or a basal porridge that had 9 micronutrients (n = 370), which they would receive daily for a year (79). Both porridges contained iron, although the dosages were different, 6.5 mg of ferrous fumarate per kilogram of porridge in the basal formulation and 250 mg of ferrous fumarate per kilogram of porridge in the rich formulation (79). At baseline the rates of iron deficiency were 19.0% and 16.0% as assessed using serum transferrin receptor in the basal and rich groups respectively and 37.2% and 32.5% using serum ferritin (79). Furthermore, at endline the adjusted odds of anaemia, iron deficiency, and iron deficiency anaemia were lower in the rich group as compared to the basal group (79). Also of note was the finding that children in households with a higher socioeconomic status (SES) at baseline saw a statistically significant greater improvement in both their haemoglobin concentrations and iron status during the course of the intervention (79).  However, given that this study was done in the capital, Lusaka, it is unlikely that the findings can be generalized to the country as whole, particularly rural regions. Furthermore, as has been the case with other studies, there is the issue of non-standard cut-offs: the authors used 28  different cut-offs for haemoglobin, serum transferrin receptor and ferritin than are typical. Finally, given that there were a large number of micronutrients in the porridges, including the basal porridge (both for example contained B12), it is not possible to point to a specific micronutrient as being responsible for the change in haemoglobin seen during the study (79). Finally, a study performed in 2004 in the rural Mpwonge District of the Copperbelt Province is notable in that it focused on children under the age of 12 months, measuring iron deficiency and feeding practices in 91 child-caregiver pairs (14). Infant haemoglobin concentrations were measured at birth, 2 months, 4 months, and 6 months of age. At birth 13% of the infants were already anaemic, this dipped to 2% at 2 months, before rising to 24% at 4 months, and 52% at 6 months (14). They also measured zinc protoporphyrin, elevated concentrations of which are marker of iron status, at birth, 4 months, and 6 months of age, and found that 9% of the infants were born iron deficient, and that by 6 months this had increased to 75%. Combined with the results for anaemia, this represented a 47% prevalence of IDA at 6 months (14). They also compared the haematological profiles of infants exclusively breastfed up to 4 months and those who started complementary feeding before 4 months, at both 4 months of age, and again at 6 months of age. At 4 months of age, infants who had already started receiving complementary foods had lower haemoglobin concentrations (p = 0.033) and a higher prevalence of anaemia (33% versus 8%, p = 0.024) (14). At 6 months there was still a significant relationship between feeding practices and haemoglobin concentration (p = 0.015), but it no longer resulted in a difference in the prevalence of anaemia (14). Thus, this study suggests not only a relationship between feeding practices and anaemia, but also that there is a high prevalence of iron deficiency by 6 months of age. 29  However, there were several problems with the study. One is the uncertainty of the cut-offs for anaemia for children under 6 months; there are currently no WHO or UNICEF guidelines for this age group. Another set of limitations is related to the decision to use zinc protoporphyrin as the marker for iron status. It is not clear that the standard cut-off for diagnosing iron deficiency is appropriate when analyzing cord blood, as was done in this study. The authors also suggest that the reported prevalence of iron deficiency may be an overestimate as zinc protoporphyrin is elevated during chronic inflammation, haemoglobinopathies, and malaria (14). This is particularly troubling given that there was no attempt to measure of the former two, and that measurements of the latter revealed a 13% parasitism rate in the 6 month old infants (14). Thus in summing up the knowledge on the causes of anaemia, there are consistent findings suggesting a link to malaria (9,74,75,78), and possibly helminthes (9). The findings on the importance of iron deficiency are mixed, and perhaps differ from region to region (14,76–79). There is also some evidence that there may be an association between socio-economic status and anaemia (79), although once again this is not consistent (9), as well as a relationship between anaemia and α-thalassaemia, although this is based on one study of limited size (59). 2.5. Measurement of Iron Deficiency Given the centrality of iron deficiency, and thus IDA, in the understanding of anaemia, it is important to understand how iron can be measured, particularly in children. While there are numerous biochemical measures to assess iron status, this review will focus only on serum ferritin, serum transferrin receptor (STfR), and their related measures: the STfR/log ferritin index, sometimes referred to as the R/F index, and the Cook Equation, used to measure body iron stores. However, before discussing these biochemical indicators, there will be a brief discussion 30  of the gold standard for measuring iron: bone marrow examination. Other biochemical indicators not discussed here include zinc protoporphyrin, transferrin saturation, and total iron binding capacity (TIBC) (31). 2.5.1. Bone Marrow Examination The “gold standard” for the determination of iron deficiency is an examination of stainable iron in a sample of aspirated bone marrow (49). However, there are a number concerns with its use, chief among them being that it is highly invasive (80). Other, more methodological concerns are that there can be bias introduced by observer error, as well as concern that stainable iron may not actually correlate well with  iron therapy (49). As such, it is rarely used in large epidemiological studies. 2.5.2. Serum Ferritin The majority of the storage iron within the body is found in the protein ferritin, which acts as a storage protein and is mostly found in the liver, as well the bone marrow and the spleen (80). However, beyond these stores there is also a small amount of ferritin found with the blood plasma, known as serum ferritin, although its function is not understood (80). Despite this, serum ferritin is considered a good indicator of body iron stores, increasing as stores increase, and decreasing as body iron stores become depleted (23). The recommended cut-off for diagnosing iron deficiency from serum ferritin depends upon the age of the individual, but for children under 5 years, the WHO cut-off is <12 µg/L and is the same for both males and females. Importantly, this cut-off is considered to have a very high specificity, as there are very few conditions other than iron deficiency that could lower serum ferritin levels (52).  31  Unfortunately, there are two issues with the use of serum ferritin to measure iron status. The first concern is that serum ferritin is unable to indicate the severity of iron deficiency from a functional standpoint, because once body iron stores are depleted, serum ferritin will stabilize, placing a floor on how low it can fall (80). This becomes most problematic when assessing the iron levels of populations that have a very high iron requirement. Examples of such groups are pregnant women and children 6-11 months old (80).  The second concern with serum ferritin is that in populations with high rates of inflammation the sensitivity of the traditional cut-off becomes very low (23). This is because serum ferritin is a positive acute phase protein. More specifically, when macrophages detect the presence of tissue damage and/or an infection, they release the cytokines interleukin 1 (IL-1) and tumor necrosis factor (TNF) to commence an inflammatory response. This results in the release of interleukin 6 (IL-6) which, along with TNF and interferon gamma (INFγ) cause an increase in serum ferritin (52). The implication of this series of events is that inflammation can mask iron deficiency if one is using traditional cut-offs. There are currently several methods to deal with the positive acute phase response of serum ferritin in children (24). The simplest is to increase the cut-off across the board for all individuals, from 12 µg/L to 30 µg/L, which is the current WHO recommendation (23). In support of this, some research has suggested the existence of an inflection point in the curve describing the relationship between haemoglobin and the log of serum ferritin. Below this point hemoglobin concentrations and the log of serum ferritin are directly related and above it there is an inverse relationship. Of note,  this inflection point corresponds roughly to the 30 µg/L cut-off for serum ferritin suggested by the WHO in the presence of inflammation, giving it a theoretical backing (81).  32  However, it is also possible to make use of other acute phase proteins to detect the presence of inflammation, as they respond to the same cytokines as serum ferritin. Typically these are C-reactive protein (CRP) and α-1-acid-glycoprotein (AGP), although α-1-anti-chymotrypsin (ACT) has been used as an alternative to CRP in some studies (52). CRP responds quickly to inflammation and falls rapidly, whereas AGP rises more slowly, but stays elevated for longer periods of time (82). Specifically, both CRP and AGP concentrations are elevated when IL-1 and TNF concentrations rise, but CRP peaks within 1 to 2 days, and has a half-life of 2 days, as opposed to the 3 to 5 days it takes AGP to peak, with a half-life of 5.2 days (52).  This allows the possibility of adjusting serum ferritin concentrations for the effect of inflammation, whereby CRP and AGP are used to split a population into four different sub-categories, based upon their stage in the acute phase response, as described by Thurnham et al. (83). These groups are a reference group with no inflammation, an incubation phase group in which only CRP is elevated, an early convalescence phase group in which both CRP and AGP are elevated, and a late convalescence phase group in which only AGP is elevated (83). While, there is some debate in the literature as to what cut-offs to use for CRP and AGP, which can make comparisons across papers difficult (82), this method uses a CRP cut-off of >5 mg/L and an AGP cut-off of >1 g/L to define inflammation (52). Within each group the mean log ferritin concentration is calculated, and the differences of the means to the mean in the reference group are back transformed, yielding the ratio of the geometric mean serum ferritin concentrations between groups. These ratios are used to create “correction factors” to adjust the serum ferritin cut-offs in each group (83). The geometric mean, as opposed to the arithmetic mean, is used due to the fact that serum ferritin values are often heavily skewed.  When applied to a meta-analysis of papers measuring the acute phase response 33  and serum ferritin, it was suggested that a new cut-off for the incubation and late convalescence groups should be 15 µg/L and that the cut-off for the early convalescence group should be 22 µg/L, although it is also possible to calculate study specific cut-offs with one’s own data (52). A variation of this method is to use the median serum ferritin concentrations when calculating the correction factors, as opposed to the geometric mean (24). There is another commonly used method for taking inflammation into account when using serum ferritin: excluding all individuals from analysis that show elevated CRP and/or AGP, then analyzing the remaining individuals using the traditional cut-off for serum ferritin (84). However, this is statistically questionable at best, as it introduces bias into the analysis, a bias that is worsened if one is forced to exclude the majority of a sample, as might be the case where there is a high infection pressure (24). 2.5.3. Serum Transferrin Receptor (STfR) Another important indicator of iron status is serum transferrin receptor (STfR), sometimes referred to as soluble transferrin receptor or plasma transferrin receptor. STfR is closely related with transferrin, the main transport protein for iron in the body (80). When iron stores are high the chemical hepcidin degrades ferroportin on the surface of cells, preventing them from releasing iron for transport to other body tissues. When there is unmet demand by tissues for iron, such as is seen during some anaemias, there is a drop in hepcidin concentrations, which facilitates transport of iron across cell membranes, where it binds with apotransferrin, producing transferrin. (85). Circulating transferrin binds with transferrin receptors located on the surface of cells, and is transported across the cell membrane, delivering iron to the cell. Afterwards, apotransferrin is released back into the blood stream, along with part of the transferrin receptor, becoming STfR.  34  The purpose of STfR in the body is not known, but because transferrin receptor is found on the surface of cells in proportion to their iron needs, it provides a good indicator of the functional iron needs or deficiency of an individual (80). This is particularly relevant because unlike serum ferritin, which will stop falling after iron stores are depleted, STfR concentrations will continue to rise as the functional iron deficiency becomes more and more pronounced (80). Unfortunately, the choice of STfR cut-offs, and even comparison of values across different studies, is complicated by the fact that there are currently three different assays for assessing the concentration of STfR (86). The most common is the so called Ramco assay from Ramco Laboratories, which makes use of an Enzyme Linked Immunoabsorbent Assay. This is followed by the Roche assay, by Roche Diagnostics, used in National Health and Nutrition Examination Survey, and the Flowers assay, used in the calculation of body iron stores. Fortunately, equations have been developed to allow the conversion of data from one assay to another (86). Furthermore, it has been suggested that as long as the appropriate cut-off is used, prevalence rates derived from the different methods should be similar (87) . When using the Ramco assay, the most commonly used cut-off for diagnosing iron deficiency in children is >8.3 mg/L (81); however, this value is not without controversy (88). A recent meta-analysis of STfR in the diagnosis of IDA found that it had a sensitivity of 86% and a specificity of 75%, although not all of the studies used in the analysis were in children or in settings of high inflammation (89). One of the major benefits of using STfR to diagnose iron deficiency is that unlike serum ferritin it is not believed to be an acute phase responder, as such it holds promise for distinguishing ACD from IDA; however it is believed to react to malaria, even when asymptomatic (24). The issue is that there is no clear consensus as to what size or even direction 35  of this effect might be (24). The destruction of red blood cells during malaria infection could induce an increase in erythropoeis and thus STfR, but there is also evidence that malaria can reduce erythropoeisis, which would lower concentrations of STfR, and it is not clear which effect is greater (24). Investigations into this effect have yielded inconsistent results, with some finding that malaria increased STfR (90,91), and some finding that malaria decreases STfR (92).  As a result of this controversy there is no accepted way to adjust STfR concentrations for the presence of a malaria infection, as has been suggested for serum ferritin and inflammation. This indicates the need for caution when interpreting STfR in the presence of endemic malaria. It has been suggested that to avoid this complication one could exclude children from analysis based upon elevated CRP and/or AGP, which would be indicative of malaria infection (90). However, as has been previously argued this would introduce bias into the results which is statistically unjustifiable. Another potential concern with the use of STfR to diagnose iron deficiency is the effect of different haemoglobinopathies on STfR concentrations. A study of 181 children from Vanuatu found that α-thalassaemia, including both the –α3.7 and –α4.2 variants, was associated with a statistically significant (p = 0.014) reduction in log STfR concentrations. The mean STfR concentration was 2.48 mg/L for non-carriers, 2.86 mg/L for heterozygous carriers, and 3.1 mg/L for homozygous carriers. From this, the authors of the study suggested caution in the use of STfR in areas with high rate of α-thalassaemia (93). A recent study of 113 individuals in Thailand found similar results, indicating that α-thalassaemia could increase STfR concentrations (94). The authors found that the mean STfR concentration for non-carriers was 0.94 mg/L, as opposed to 1.06 mg/L for α-thalassaemia heterozygotes of the variant also found in those of African descent (94).  Similar results have also been reported when examining the effect of sickle cell 36  disease on STfR concentrations (95). Unfortunately, as is the case with malaria, there is currently no way to correct for the effect of these conditions on concentrations of STfR. 2.5.4. Serum transferrin (STfR), Log Ferritin Index, and the Cook Equation In recognition of the need for a test of iron deficiency that has high sensitivity even in the presence of infection, some researchers have suggested a combined measure, the STfR/log ferritin index, also referred to as the R/F index, or TFR-F index (96). The concept behind this index is that by combining information from STfR, which is reflective of functional iron needs, and serum ferritin, which is reflective of body iron stores, one can achieve higher specificity and sensitivity than either measure alone (96), although this has been contested in a recent meta-analysis, which found that elevated STfR was better at differentiating IDA and ACD (89).  A recent study in Tanzania, comparing measures of iron deficiency in children against the gold standard, bone marrow examination, in the presence of disease found that the STfR/log ferritin index, when using a cut-off of >5.6, had a sensitivity and specificity of 70% and 75% respectively (88). This compared favourably to the WHO ferritin cut-off of <30 µg/L, which had a high specificity of 96%, but a low sensitivity of 21% and the STfR cut-off of >8.3 mg/L which had a high sensitivity of 90%, but a low specificity at 37% (88). Furthermore, if one adjusted the STfR/log ferritin cut-off to >5.3 the sensitivity and specificity became 74% and 73% respectively (88). The key strengths of this paper were that it was in a sample with a high prevalence of inflammation, and that they were comparing values against the gold standard, stainable iron in the bone marrow (88). Another combined measure is the so-called Cook Equation, which attempts to quantify body iron stores (BIS) using both serum ferritin and STfR (97). This is sometimes referred to as the STfR:SF ratio, but it is distinct from the STfR/log ferritin index. The Cook Equation is an 37  attempt to quantify the actual size of the body iron pool, in milligram per kilogram body weight; a positive value indicates the amount of iron in the body, whereas a negative number indicates the amount of iron that would need to be absorbed to correct the current deficiency (80). As such the cut-off for diagnosing iron deficiency is <0.0 mg/kg iron, which makes for an easier interpretation of values (80). While the Cook Equation was experimentally derived, its major limitation in the present research is that it has not been validated for use with children (97). 2.6. Summary of Key Gaps in the Literature From the literature review above, there are a number of key gaps which this thesis will help to address. Anaemia is clearly a serious condition which affects a large portion of the population; however, its precise etiologies differs from place to place. While there has been some work in Zambia on this problem, it is still fairly limited, a gap that this thesis will help fill. One example of this is the association between distal or household variables, such as maternal education, on anaemia. Another gap this thesis will explore is the association between IYCF indicators, such as minimum dietary diversity, minimum meal frequency, and minimum acceptable diet and anaemia as well as breastfeeding practices. This thesis may also serve to confirm previous findings, such as the lack of an association between the use of malaria nets and anaemia in the Zambian context. In looking at the biochemical and household factors associated with anaemia and haemoglobin concentrations, this thesis will be helpful not only in designing future experimental studies, necessary to establish causal pathways, but will also be able to provide critical information for the design of interventions to reduce rates of anaemia in the country. There is also a lack of information on iron deficiency in the Zambian context, which is especially troubling given the high rates of anaemia, and the assumed, though unconfirmed, 38  connection between the two. This thesis will help to fill this gap by providing estimates of iron deficiency using different indicators, as well as exploring the association between iron status and anaemia. Because this study measured multiple indicators of iron status alongside measures of inflammation, it will be possible to contribute to the literature analyzing the relationship between these iron status indicators in the context of inflammation, a gap which at the moment makes it very difficult to state the prevalence of iron deficiency in some contexts with confidence.   39  3. Methods  This study was implemented by the Ministry of Health (MoH) and Ministry of Community Development, Mother, and Child Health (MCDMCH) in partnership with UNICEF-Zambia with funds from Irish Aid. The University of British Columbia provided technical assistance to the project.  The thesis represents analyses of baseline data from an effectiveness study examining the effect of MNP, paired with IYCF training, on anaemia and other biomarkers of nutrition, which was nested within a larger pilot project of MNP distribution in Mbala District, Northern Province. Within the effectiveness study, there were two treatment arms: a control group and a treatment group. Both groups received IYCF, de-worming tablets (Albendazole) and insecticide-treated mosquito nets, while the treatment group additionally received MNP. The baseline data collection took place before any MNP, training, de-worming tablets, or mosquito nets were delivered. For the rest of this section, I will focus on the methods for the baseline evaluation of the effectiveness study, as the later midline and endline evaluations are not part of this thesis. 3.1. Sampling and Participants The effectiveness study made use of a convenience sample. This was based on feedback from the local partners at UNICEF-Zambia and the Zambian Ministry of Health. Given the sensitive nature of blood collection in the local context, it was decided that randomized selection would not be feasible, and would likely result in low acceptance and enrolment. As the effectiveness study is nested within the larger pilot project, in which an even larger group of mothers would be given MNP but not followed, sampling was based around Catchment Areas used by the local health system, and through which the MNP were to be distributed. Of the total 26 Catchment Areas in Mbala District, five Catchment Areas were selected for the treatment arm 40  and four Catchment Areas for the control arm. The treatment Catchment Areas were Kawimbe, Mbala Urban, Tulemane, Mpande, and Mambwe Mission, whereas the control Catchment Areas were Kamuzwazi, Kaka, Senga and Nondo. Catchment Areas were selected purposely, to ensure a mix of both rural and more urbanized (such as Mbala Urban and Tulemane) areas, as well as a mix of areas both close to, and more distant from, the main paved road in the district. Catchment Areas which were too close to the Tanzanian border were excluded. Given how porous it is, people often cross back and forth, making follow up more difficult. Furthermore, some Catchment Areas were excluded based upon feedback from local partners about specific areas which were likely to refuse the intervention for religious or cultural reasons. A map of the selected Catchment Areas is shown in Figure 3.1. 41  Figure 3.1 Map Showing Catchment Areas in Mbala District, Northern Province, Zambia, Which Were Selected for Sampling.  Catchment Areas in Zambia are further split into zones, each with either a full health clinic or a much smaller health outpost, which are sometimes located in a school. Within each of the selected Catchment Areas, five or six zones were selected for sampling. Zones not accessible 42  during the rainy season due to the potential for poor roads and with very low populations were excluded. A complete list of the selected zones is included in Appendix A. 3.2. Ethical Consideration Ethical approval was obtained from both the Tropical Disease Research Centre Ethics Review committee in Zambia, approval number TRC/C4/05/2013, and the University of British Columbia’s Clinical Research Ethics Board, approval number H13-00261. The study was registered with ClinicalTrials.gov as NCT01878734. Participation of individuals in the study was voluntary. Participation could be withdrawn at any time without consequences, and informed consent was obtained from caregivers of children, with the project being explained to them in their own language. Consent was always obtained in the presence of a local community health worker or community health volunteer, who would act as a witness. The consent forms are included in Appendix B. 3.3. Inclusion and Exclusion Criteria  The inclusion and exclusion critieria are listed below.  Inclusion Criteria: i. Children aged 6-11 months; to ensure children will be within the eligible age group of 6-23 month for the duration of the entire pilot;  ii. Children residing within one of the 9 project Catchment Areas as defined by the study, and who plan on remaining in the same household for the 12 month study duration; iii. Parent/guardian willingness to give consent for the child’s participation in the study/   43  Exclusion Criteria: i. Severely malnourished children (weight-for-height Z-score < -3 SD, MUAC < 11.5 cm, and/or the presence of bilateral oedema); ii. Children with severe anaemia (haemoglobin concentration < 70 g/L);  iii. HIV positive children, as determined by their Zambian ‘Under 5 Child Card’, any other documentation, or during the baseline questionnaire. 3.4. Evaluation Tools Three broad categories of data were collected during the baseline assessment: household and other information was obtained by means of questionnaire; anthropometry on all children; and biochemical data. 3.4.1. The Questionnaire A baseline questionnaire was administered to each caregiver by a Zambian enumerator. The questionnaire was informed by previous formative research in the country and contained seven modules with the following categories: ‘Anthropometry, Oedema & Biochemical Assessment’ which was used to record the anthropometry and haemoglobin values; ‘Household Information’ which gathered information on age, number of children, demographics and livelihoods; ‘Drinking Water, Hygiene & Sanitation’ which gathered information on hand-washing, water sources and purification, and availability of toilet facilities; ‘Early Childhood Development’ which measured caregiver knowledge and practices in terms of active feeding and leaving children unattended; ‘Child Health & Health Seeking Behaviour’ which measured caregiver reported symptoms of morbidity, and how they responded when a child was ill; ’Caregiver Knowledge of Infant and Young Child Feeding Practices  & Anaemia’ which measured caregiver perceptions, and ‘Infant and Young Child Feeding Practices & Food 44  Consumption’ which measured actual practice, as well as determining the child’s food consumption over the previous 24 hours using a check list.  This information was used to calculate the standard IYCF indicators of Minimum Meal Frequency, Minimum Dietary Diversity, and Minimum Acceptable Diet, using definitions laid out by the WHO (44). Minimum Meal Frequency is defined as consuming 2 or more meals per day in breastfed children 6-8 months, 3 or more meals per day in breastfed children 9-23 months, and 4 or more meals per day in non-breastfed children 6-23 months. Minimum Dietary Diversity is defined as consuming four or more of the seven following food groups: grains, roots and tubers, legumes and nuts, dairy products, flesh foods, eggs, vitamin rich fruits and vegetables, and other fruits and vegetables. Minimum Acceptable Diet is the met in children achieving both Minimum Meal Frequency and Minimum Dietary Diversity for breastfed children, and those achieving 2 milk feedings, Minimum Meal Frequency not including the milk feedings, and Minimum Dietary Diversity for non-breastfed children. The questionnaire was administered in one of two local languages, Bemba or Mambwe, depending on the preference of the caregiver as many did not speak English. Interviewers made use of pre-translated questionnaires. However, all responses were recorded in English. Prior to being used in the field the questionnaire was piloted at the Kalingalinga Health Centre in Lusaka with local mothers and their children. Based on this experience, the questionnaire was modified to change any questions that were found to be problematic to code or confusing for the mothers. A copy of the questionnaire is included in Appendix C for reference. 45  3.4.2. Anthropometric Data The length, weight, age and mid upper-arm circumference of each child was obtained in order to calculate z-scores for weight-for-age (underweight), height-for-age (stunting) and weight-for-height (wasting), as derived from the WHO growth curves.  Length was measured using a standard wooden recumbent length board, as children were too young to stand. Length was measured twice, to the nearest 1mm, and averaged. If the two measurements disagreed by more than 5mm a third measurement was taken and the two closest values were used. Hats, socks, shoes, and baggy clothes were removed from the children before their length was measured and the child was removed from the length board in between measurements.  Weight was measured using a battery powered digital flat scale manufactured by SECA™. Caregivers would first stand on the scale with shoes removed, allowing it to be tared, before being handed their child. Excess clothing was first removed from the child. The weight was then recorded to the nearest 100 grams. The scales were calibrated using known weights of 2 kilograms. Measurements were always taken on level surfaces. Mid upper-arm circumference (MUAC) was measured using a standard MUAC measuring tape.  Two measurements were taken to the nearest 1 mm, and averaged. If the two measurements disagreed by more than 5mm a third measurement was taken and the two closest values were used. Shirts were removed from children before being measured. Age was measured by determining the child’s date of birth as listed on the Under-5 Card, and the use of an Age Calculator. In instances where the caregiver did not have an Under-5 Card, 46  the date of birth recorded in other documentation, such as the notebook used by health workers to record vaccinations and weights, was used. 3.4.3. Blood Collection A non-fasting blood sample was collected by venpuncture into evacuated tubes containing EDTA. Haemoglobin concentration was measured on the spot using a portable photometer, the Hemocue™, using a drop of venous blood. A slide was prepared with a thick blood smear of venous blood to diagnose malaria.   Anaemia was defined as haemoglobin (Hb) < 110 g/L, with 3 levels of severity: mild anaemia Hb 100-109 g/L, moderate anaemia Hb 70-99 g/L, and severe anaemia Hb < 70 g/L (29). Presence of malaria was determined by use of microscopy of the thick blood smears, which were analysed by technicians at the Tropical Disease Research Centre (TDRC) in Ndola, Lusaka, resulting in counts of both parasites and gametocytes per 200 white blood cells. The remaining blood was centrifuged and the serum removed, aliquoted, and stored at -20 C until analysis for ferritin, serum transferrin receptor (sTfR), retinal binding protein (RBP), AGP (alpha-1 acid glycoprotein ) and CRP (C-reactive protein) using a sandwich enzyme-linked immunosorbent (ELISA) at the Erhardt Laboratory in Germany (50). Serum retinol-binding protein (RBP) is a marker of vitamin A status and as is considered to be more heat and light stable than serum retinol, which makes it ideal for field studies (98). CRP (C-reactive protein) and AGP (alpha-1 acid glycoprotein), both acute phase proteins,  were used as biomarkers of underlying infections (52). CRP provides an indicator of acute disease while AGP is a marker of chronic infections (52). Iron status was determined by measuring serum ferritin and STfR (49).  47  Vitamin A deficiency was defined as RBP < 0.7 µmol/L (98). CRP was considered elevated at values > 5 mg/L, and AGP was considered elevated at values >1 g/L (52). An individual with neither elevated was considered not to have inflammation, an individual with only CRP elevated was considered to be in the incubation phase of the infection cycle, an individual with both CRP and AGP elevated was considered to be in the early convalescence phase, and an individual with only AGP elevated was considered to be in the late convalescence phase (52). Three different classifications were used to assess iron deficiency based on serum ferritin. The first made use of serum ferritin < 30 µg/L, elevated from the typical cut-off for this age group of < 12 µg/L  due to the high burden of disease (99). The second made use of CRP and AGP to split individuals into the four inflammation phases, or groups, with each phase having its own serum ferritin cut-off to diagnose iron deficiency. There were: Serum ferritin < 12 µg/L, serum ferritin < 15 µg/L, serum ferritin < 15 µg/L, and serum ferritin < 22 µg/L were used for the no inflammation (CRP ≤ 5 mg/L and AGP ≤ 1 g/L), incubation (CRP > 5 mg/L and AGP ≤ 1 g/L), late convalescence (CRP ≤ 5 mg/L and AGP > 1 g/L) and early convalescence (CRP > 5 mg/L and AGP > 1 g/L) groups respectively (52). A third definition of iron deficiency used serum ferritin < 12 µg/L, but excluded anyone that had signs of inflammation from either CRP or AGP (99). The three different definitions of iron deficiency that used serum ferritin are summarized below in Table 3.1.  48  Table 3.1  Serum ferritin cut-offs used to diagnose iron deficiency in for each inflammation phase, as determined by C-reactive protein and alpha-1-acid glycoprotein concentrations Serum Ferritin Classification No Inflammation Incubation Phase Early Convalescence Phase Late Convalescence Phase WHO elevated cut-off Serum ferritin < 30 µg/L Serum ferritin < 30 µg/L Serum ferritin < 30 µg/L Serum ferritin < 30 µg/L Inflammation adjusted cut-offs Serum ferritin < 12 µg/L Serum ferritin < 15 µg/L Serum ferritin < 22 µg/L Serum ferritin < 15 µg/L Exclusion method Serum ferritin < 12 µg/L Excluded from analysis Excluded from analysis Excluded from analysis  A concentration of STfR > 8.3 mg/L was also used to define iron deficiency (24). Iron deficiency anaemia was defined as concurrent iron deficiency and anaemia, using the above definitions. Coefficient of variation was calculated using known controls from the CDC and Biorad Liquicheck for serum ferritin, STfR, RBP, CRP and AGP. The corresponding coefficients of variation were: 3.20%, 3.03%, 3.76%, 5.18% and 5.11%. 3.5. Data Analysis and Statistics All data was entered on the day of collection using an Excel™ spread-sheet. Cleaning of data took place during data entry. Validation rules embedded within the Excel™ sheet prevented entry of invalid data. Data that was coded incorrectly by enumerators or not recorded at all was coded as missing data. All data was then double checked, with one person reading the questionnaire aloud, while another compared it against the already entered data. 49  After data entry and clean up, the data was transferred into an IBM SPSS file version 22.0.0.0 for analysis. The HC3 standard error estimator described below made use of an SPSS macro developed and published by Hayes and Cai (100). Anthropometric calculations were created using WHO Anthro version 3.2.2, and then also transferred into SPSS. Descriptive statistics made use of means and percentages, calculated using valid N, as opposed to total N. Bivariate analysis made use of independent samples t-tests, chi-square tests, and Spearman’s correlations where appropriate. A p-value < 0.05 was used to determine significance, and associations with a p-value < 0.2 were entered into regression models. The final decision on whether to include or drop a given predictor was based upon the significance of its β and knowledge on the etiology of anaemia, both in Zambia and elsewhere. Logistic regression was used to create an explanatory model of anaemia status, with models assessed using the Hosmer and Lemeshow goodness of fit test and the Nagelkerke R Square. Sensitivity and specificity of the models were also calculated. The assumption of linearity of independent continuous variables relative to the logit was tested using the Box-Tidwell procedure with a Bonferroni correction (101). The significance of individual terms was checked by examining the Wald statistic. In constructing the model, only variables which had a significance of p < 0.2 in bivariate analysis were included, with terms removed until all were significant at p <0.05, producing a more parsimonious model. Multiple linear regression was used to create an explanatory model of haemoglobin concentration. The assumption of independence of observations was tested using the Durbin-Watson statistic. The assumption of linearity was tested by plotting standardized residuals against unstandardized predicted values, as well as by examining partial regression plots. The assumption of homoscedasticity was tested by plotting standardized residuals against 50  unstandardized predicted values. Given the presence of heteroscedasticity, a heteroscedasticity-consistent standard error estimator, the HC3, was used when calculating p-values for individuals terms in the model, as well as confidence intervals for all β values (100). Multicollinearity was tested for by examining Tolerance values, with a cut-off of Tolerance < 0.1 used to indicate the presence of collinearity. Extreme outliers were detected and examined by flagging standardized residuals > ±3, leverage points were identified by using a cut-off of leverage > 0.5, and influence points were identified by examining Cook’s Distance values. A normal P-P plot of the standardized residuals was examined to determine whether the residuals were normally distributed. Once these assumptions were checked, model fit was assessed by examining the adjusted R Square value, the log likelihood value and the F-ratio. In constructing the model, only variables which had a significance of p < 0.2 in bivariate analysis were included, with terms removed until all were significant at p <0.05, producing a more parsimonious model.  51  4. Results 4.1. Recruitment 631 eligible caregiver-child pairs were recruited for the study and completed questionnaires. Haemoglobin measurements were obtained from 631 children, and valid blood samples for CRP, AGP, RBP, STfR, and serum ferritin were obtained from 620 children. Thick blood smears for the testing of malaria parasitism were obtained from 625 children. Valid results were obtained for 628 individuals for underweight, 629 individuals for stunting, and 627 individuals for wasting. Chi-squared tests were used to test for the possibility of terminal digit bias, and while results were not significant for the weight measurements, they were for the length and MUAC measurements (p < 0.1), suggesting that enumerators were rounding values when taking measurements. 4.2. Household and Participant Characteristics Table 4.1 shows household and participant characteristics. Approximately half of the children were female, and the mean age was 9 (SD 2) months. The average age of the primary caregiver was 27 (SD 8) years, and on average households had 3 children under 5 years residing in it. Nearly four fifths of caregivers had achieved some level of education, with the majority having completed primary schooling only. Nearly all (91%) households reported that they had some source of income, 97% had land for cultivating crops, 53% had a fruit and/or vegetable garden, and 82% owned a farm animal of some kind. Just over one third of caregivers reported that they treated their water, with the most common method being the addition of chlorine or bleach.  52  Table 4.1 Household and participant characteristics Characteristic Value Total N Female, % (n) 49.5% (312) 630 Age of child, months, mean (SD) 8.9 (1.7) 631 Age of primary caregiver in years, mean (SD) 26.7 (7.6) 631 Household size, mean (SD) 5.9 (2.4) 630 Number of children under 5 in household, mean (SD) 3.3 (2.2) 614 Education level of primary caregiver        No school, % (n) 21.7% (137) 631      Primary (Grade 1-7), % (n) 59.3% (374) 631      Secondary (Grade 8-12), % (n) 17.7% (112) 631      Tertiary (University or higher), % (n) 0.5% (3) 631      Education level not specified, % (n) 0.8% (5) 631     Household has source of income, % (n) 90.6% (569) 628 Household has land for cultivating crops, % (n) 96.5% (609) 631 Household has fruit and/or vegetable garden, % (n) 52.6% (332) 630 Household has livestock/farm animals, % (n) 81.6% (511) 626     Household treats water to make it safer, % (n) 36.0% (227) 631      Household boils water, % (n) 11.7% (74) 631      Household treats water with bleach/chlorine, % (n) 22.6% (149) 631      Household treats water, method not specified, %(n) 0.6% (4) 631 Adults wash hands with soap or ash, % (n) 86.5% (545) 630  4.3. Morbidity and Health Characteristics Table 4.2 shows key morbidity and health characteristics of the children. Caregivers reported that the 56% of the children had diarrhea, 72% a cough, and 56% a fever over the previous two weeks.  Over two thirds (69%) of children were sleeping under a mosquito net, and 10% tested positive for malaria parasites. Based upon elevated concentrations of CRP and AGP, 74% of the children were experiencing some form of inflammation. Just under three quarters (74%) had received a vitamin A capsule within the past 6 months.  53  Table 4.2 Morbidity and health of children 6-11 months, in Mbala District Morbidity/health Characteristic Value Total N Diarrhea in the past 2 weeks1, % (n) 56.1% (354) 631 Cough in the past 2 weeks, % (n) 72.1% (455) 631 Fever in the past 2 weeks, % (n) 55.6% (349) 628 More than one of the above in the past 2 weeks, % (n) 63.9% (401) 628 None of the above in the past 2 weeks, % (n) 12.6% (79) 628     Children sleeping under a mosquito net, % (n) 68.9% (433) 628 Children with fever who were tested for malaria, % (n) 46.6% (129) 277 Children who are currently being treated for malaria, % (n) 9.2% (58) 631 Children with malaria parasitism, % (n) 9.9% (62) 625 Malaria parasites per 200 white blood cells, mean (range)  23.6 (0-4000) 625    Inflammation status based on CRP2 and AGP3 concentrations        No inflammation4, % (n) 26.0% (161) 620      Only CRP elevated  (Incubation phase)5, % (n) 2.1% (13) 620      Both CRP and AGP elevated (Early convalescence phase)6, % (n) 37.1% (230) 620      Only AGP elevated (Late convalescence phase)7, % (n) 34.8% (216) 620    Given a vitamin A capsule within past 6 months, % (n) 73.7% (465) 631 Given a de-worming capsule within past 6 months, % (n) 10.3% (65) 631 1 Defined as 3 or more times a day of loose stools    2 C-reactive protein    3 Alpha-1-acid glycoprotein    2 Defined as CRP < 5 mg/L and AGP < 1 g/L    3 Defined as CRP > 5 mg/L and AGP < 1 g/L    4 Defined as CRP > 5 mg/L and AGP > 1 g/L    5 Defined as CRP < 5 mg/L and AGP > 1 g/L     4.4. Infant and Young Child Feeding (IYCF) Knowledge and Practices IYCF knowledge and practices are reported in Table 4.3. Key findings were that every child had been breastfed at some point, and that breastfeeding had only been stopped for one child. Furthermore, the average age for the commencement of complementary feeding was 6 months, which had commenced for 97% of the children. Unfortunately, only 22%, 56% and 18% of children were meeting minimum dietary diversity, minimum meal frequency, or minimum 54  acceptable diet respectively. Furthermore 28% had consumed an iron-rich food in the previous 24 hours. Table 4.3 Infant and young child feeding (IYCF) knowledge and practices of caregivers in Mbala District IYCF knowledge/practice Value Total N Child currently breastfed, % (n) 99.8% (630) 631 Child ever breastfed, % (n) 100% (631) 631 Child has commenced complementary feeding, % (n) 96.5% (609) 631 Age of start of complementary feeding, months, mean (SD) 5.9 (1.1) 609     Child meeting minimum dietary diversity1 , % (n) 22.0% (139) 631 Child meeting minimum meal frequency2, % (n) 56.1% (352) 628 Child meeting minimum acceptable diet3, % (n) 18.0% (113) 628 Child consuming iron-rich food in previous 24 hours4, % (n) 28.2 (178) 631     Caregiver heard of anaemia and/or iron deficiency, % (n) 87.0 (549) 631 Caregiver spoken to by healthcare workers about IYCF,  % (n) 69.9% (441) 631 1 Defined as consuming at least 4 of grains, roots and tubers; legumes and nuts;  dairy products; flesh foods; eggs; vitamin A rich fruits and vegetables; or other fruits and vegetables 2 Defined as 2 meals or snacks for breastfed infants 6–8 months, 3 meals or snacks for breastfed children 9–23 months, and 4 meals or snacks for non-breastfed children 6–23 months 3 Defined as at least 2 milk feedings and the minimum dietary diversity not including milk feeds and the minimum meal frequency for non-breastfed children 6-23 months and as minimum dietary diversity and minimum meal frequency for breastfed infants 6-23 months 4 Defined as flesh foods, commercially fortified foods specially designed for infants and young children that contain iron 4.5. Anthropometry Table 4.4 presents the data on stunting, wasting, underweight and MUAC for children recruited into the study. Overall, 30% were stunted, 2% were wasted, and 16% were underweight. While there were no instances of severe wasting in the sample due to the exclusion criteria, 8% were severely stunted and 2% were severely underweight. Based on MUAC, 9% of the children were suffering from moderate acute malnutrition; due to exclusion criteria there were no children with MUAC < 11.5 cm in the sample. Boys were more likely to be stunted, severely stunted, and underweight than girls, and less likely to be suffering from moderate acute 55  malnutrition as assessed by MUAC. There was no difference in the prevalence of wasting according to sex.56  Table 4.4 Stunting, wasting, underweight and mid-upper arm circumference (MUAC) of male and female children 6-11 months, in Mbala District Anthropometric characteristic Male Female p-value1 Overall Stunted (height-for-age z-score < -2), % (n) 35.6% (113) 24.7% (77) 0.004 30.2% (190)    Severely stunted (height-for-age z-score < -3), % (n) 10.7% (34) 4.5% (14) 0.005 7.6% (48) Height-for-age z-score, mean (SD) -1.7 (1.1) -1.4 (1.0) 0.002 -1.5 (1.1)      Wasted (weight-for-height z-score < -2), % (n) 2.5% (8) 2.3% (7) 1 2.4% (15) Weight-for-height z-score, mean (SD) -0.1 (1.1) 0.0 (1.0) 0.155 0.0 (1.0)      Underweight (weight-for-age z-score < -2), % (n) 21.1% (67) 10.9% (34) 0.001 16.1% (101)    Severely underweight (weight-for-age z-score < -3), % (n) 2.2% (7) 1.6% (5) 0.796 1.9% (12) Weight-for-age z-score, mean (SD) -1.0 (1.1) -0.8 (1.0) 0.005 -0.9 (1.1)      Moderate acute malnutrition (MUAC 11.5-12.5 cm) , % (n) 6.6% (21) 11.5 (36) 0.043 9.0% (57) MUAC (Mid upper-arm circumference), mean (SD) 13.9 (1.0) 13.6 (1.0) <0.001 13.8 (1.0) 1 Comparing male and female children, Chi-square test used for categorical variables and independent sample t-test for continuous 57  4.6. Vitamin A Deficiency, Anaemia, Iron Deficiency, and Iron Deficiency Anaemia Using a definition of vitamin A deficiency of RBP < 0.7 µmol/L, 5% of children were vitamin A deficient. Almost 60% of the children were anaemic, 25% mildly and 33% moderately anaemic (Table 4.5). There were no cases of severe anaemia as a result of the exclusion criteria for the study. Boys were more likely to be anaemic than girls, which was driven by a higher prevalence of moderate, as opposed to mild, anaemia. The overall mean Hb concentration was 105 g/L (SD 15). The mean Hb concentration was 103 g/L (SD 15) for boys and 107 (SD 15) g/L for girls. Table 4.5 Haemoglobin (Hb) and anaemia of children 6-11 months, in Mbala District Characteristic Male (N=318) Female (N=312) p-value1 Overall Anaemic (Hb < 110 g/L), n (%) 61.8% (196) 53.2% (166) 0.039 57.4% (362) Mildly anaemic (Hb 100-109 g/L), % (n) 23.6% (75) 26.0% (81) 0.549 24.7% (156) Moderately anaemic (Hb 70-99 g/L), % (n) 38.1% (121) 27.2% (85) 0.005 32.6% (206)      Hb concentration (g/L), mean (SD) 103.3 (15.3) 107.1 (14.6) 0.002 105.2 (15.2) 1 Comparing male and female children, Chi-square test used for categorical variables and independent sample t-test for continuous Table 4.6 shows the results of testing for iron deficiency, using four different definitions. The lowest prevalence of iron deficiency, 13%, was found when using the inflammation adjusted serum ferritin cut-offs. The highest prevalence of iron deficiency, 93% was found when using STfR > 8.3 mg/L. Usng a serum ferritin cut-off of <30 µg/L which resulted in a prevalence of 42% and the inflammation exclusion method, which used serum ferritin < 12 µg/L, while excluding all cases with evidence, and resulted in a 14% prevalence of iron deficiency. Irrespective of the definition, the prevalence of iron deficiency was always higher in boys than girls, although this relationship was not significant for the inflammation exclusion method. 58  Table 4.6 Iron deficiency in children 6-11 months, in Mbala District, using different definitions Iron deficiency definition Male (N=312) Female (N=307) p-value1 Overall Serum ferritin < 30 µg/L 47.8% (149) 35.5% (109) 0.003 41.8% (259) Inflammation adjusted ferritin cut-offs2 18.9% (59) 7.2% (22) <0.001 13.2% (82) Serum ferritin < 12 µg/L, excluding inflammation3 19.8% (16) 8.8% (7) 0.077 14.3% (23) Serum transferrin receptor (STfR) > 8.3 mg/L 96.2% (300) 89.3% (274) 0.002 92.7% (575) 1 Comparing male and female children, Chi-square test used  2 Defined as serum ferritin < 12 µg/L for no inflammation, serum ferritin < 15 µg/L for incubation and late convalescence, and serum ferritin < 22 µg/L early convalescence groups from inflammation markers 3 Defined as alpha-1-acid glycoprotein (AGP) > 1 g/L and/or c-reactive protein (CRP) > 5 mg/L   Table 4.7 shows the prevalence of IDA as determined by using the four definitions of iron deficiency. The results mirror those for iron deficiency, with the lowest prevalence of IDA, 8%, found when using inflammation adjusted serum ferritin cut-offs to determine iron deficiency. The highest prevalence of IDA, 53%, was found when using STfR to diagnose iron deficiency. As with iron deficiency, boys have a higher prevalence of IDA regardless of the definition of iron deficiency; again, this relationship was not significant for the inflammation exclusion method. 59  Table 4.7 Iron deficiency anaemia (IDA) in children 6-11 months, in Mbala District, using different definitions Iron deficiency definition Male (N=312) Female (N=307) p-value1 Overall Serum ferritin < 30 µg/L 26.9% (84) 17.3% (53) 0.005 22.1% (137) Inflammation adjusted ferritin cut-offs2 11.5% (36) 4.9% (15) 0.004 8.2% (51) Serum ferritin < 12 µg/L, excluding inflammation3 12.3% (10) 6.3% (5) 0.289 9.3% (15) Serum transferrin receptor (STfR) > 8.3 mg/L 59.0% (184) 47.2% (145) 0.004 53.1% (329) 1 Comparing male and female children, Chi-square test used 2 Defined as serum ferritin < 12 µg/L for no inflammation, serum ferritin < 15 µg/L for incubation and late convalescence, and serum ferritin < 22 µg/L early convalescence groups from inflammation markers 3 Defined as alpha-1-acid glycoprotein (AGP) > 1 g/L and/or c-reactive protein (CRP) > 5 mg/L   4.7. Bivariate Analysis of the Factors Associated with Haemoglobin Concentration There was a statistically significant (p < 0.05) positive relationship between haemoglobin concentration and the age of the child (r = 0.10), the stunting z-score (r = 0.09), the underweight z-score (r = 0.09), and the mid-upper arm circumference (r = 0.10) (Table 4.8). There was a statistically significant (p < 0.05) negative relationship between haemoglobin concentration and the number of malaria parasites in the blood of a child (r = -0.25), serum ferritin concentration (r = -0.11), STfR concentration (r = -0.33), CRP concentration (r = -0.24), and AGP concentration (r = -0.23). There was no relationship between inflammation adjusted serum ferritin concentrations and haemoglobin concentrations. 60  Table 4.8 Bivariate comparisons of haemoglobin concentration against continuous variables in a sample of 6-11 month old infants in Mbala District Variable Spearman’s correlation p-value Total N Age of child (months) 0.10 0.016* 631 Age of caregiver (years) 0.05 0.190 631 Number of individuals in household 0.02 0.622 630 Number of children <5 years in household <0.01 0.955 614 Malaria parasites per 200 white blood cells -0.25 <0.001* 625 Month at which complementary feeding began 0.05 0.239 609 Wasting z-score 0.04 0.341 627 Stunting z-score 0.09 0.032* 629 Underweight z-score 0.09 0.028* 628 Mid-upper arm circumference (cm) 0.10 0.014* 631 Serum ferritin concentration (µg/L) -0.11 <0.007* 620 Inflammation adjusted serum ferritin concentration (µg/L)1 -0.07 0.064 620 Serum transferrin receptor concentration (mg/L) -0.33 <0.001* 620 C-reactive protein concentration (mg/L) -0.24 <0.001* 620 Alpha-1-acid glycoprotein concentration (g/L) -0.23 <0.001* 620 Retinol binding protein concentration (µmol/L) -0.02 0.633 620 1 Defined as serum ferritin concentrations from the incubation, early and late convalescence groups, as defined by concentration of C-reactive protein and alpha-1-acid glycoprotein, multiplied by 0.77, 0.53 and 0.75 respectively  * Significant at p < 0.05 As shown in Table 4.9, there was a statistically significant (p < 0.05) relationship with haemoglobin concentration for sex and whether or not a household treated their water. Males had a lower average haemoglobin concentration as did households that did not treat their water. There was no statistically significant relationship between haemoglobin status and whether the caregiver had received formal education. 61  Table 4.9 Bivariate comparisons of haemoglobin (Hb) concentration against categorical household and participant characteristics in a sample of 6-11 month old infants in Mbala District Characteristic Mean Hb (g/L) SD p-value1 Male 103.3 15.3 0.002* Female 107.1 14.6     Caregiver has attended school 105.4 15.3 0.511 Caregiver has not attended school 104.4 14.8     Household has source of income 105.5 15.3 0.136 Household has no source of income 102.4 14.0     Household has land for cultivation 105.1 15.2 0.762 Household has no land for cultivation 106.1 15.6     Household has garden for fruit and/or vegetables 105.7 15.9 0.365 Household has no garden for fruit and/or vegetables 104.6 14.4     Household owns animals or livestock 105.3 15.4 0.634 Household owns no animals or livestock 104.6 14.7     Household treats water 107.0 15.7 0.021* Household does not treat water 104.1 14.8     Caregiver washes hands with soap or ash 105.5 15.4 0.24 Caregiver does not wash hands with soap or ash 103.4 13.7  1 Calculated from comparing means using an independent samples t-test * Significant at p < 0.05 As reported in Table 4.10, there was a statistically significant (p < 0.05) relationship between haemoglobin concentration and whether or not the child had had diarrhea or a fever in the previous two weeks. Children who had had diarrhea saw a 2.5 g/L reduction in haemoglobin concentrations compared to those who did not have diarrhea, and children who had had a fever saw a 3.4 g/L reduction in haemoglobin concentrations. Of note, there was no relationship between haemoglobin concentration and whether or not a child slept under a mosquito net. 62  Table 4.10 Bivariate comparisons of haemoglobin (Hb) concentration against categorical morbidity and health characteristics in a sample of 6-11 month old infants in Mbala District Characteristic Mean Hb (g/L) SD p-value1 Child diarrhea in the previous two weeks 104.1 15.6 0.042* Child diarrhea free in the previous two weeks 106.6 14.5     Child had a cough in the previous two weeks 104.9 15.3 0.421 Child cough free in the previous two weeks 106.0 14.9     Child fever in the previous two weeks 103.6 15.3 0.005* Child fever free in the previous two weeks 107.0 14.9     Child sleeps under a mosquito net 105.3 15 0.626 Child does not sleep under a mosquito net 104.7 15.6     Child has received a vitamin A capsule in the previous 6 months 105.5 15.3 0.249 Child has not received a vitamin A capsule in the previous 6 months 103.9 14.9     Child has received a deworming tablet in the previous 6 months 107.7 16.6 0.147 Child has not received a deworming tablet in the previous 6 months 104.8 15 1 Calculated from comparing means using an independent samples t-test * Significant at p < 0.05 Table 4.11 shows the associations between haemoglobin concentrations and categorical anthropometric and biochemical variables. Children with moderate acute malnutrition, as assessed by MUAC, had lower average haemoglobin concentrations, as did children who were iron deficient using STfR the inflammation exclusion method. There was not a significant difference in mean haemoglobin concentrations for the other methods of assessing iron deficiency. 63  Table 4.11 Bivariate comparisons of haemoglobin (Hb) concentration against categorical anthropometric and biochemical characteristics in a sample of 6-11 month old infants in Mbala District Characteristic Mean Hb (g/L) SD p-value1 Stunted (height-for-age z-score < -2) 103.5 15.1 0.068 Not Stunted (height-for-age z-score ≥ -2) 105.9 15.2     Wasted (weight-for-height z-score < -2) 104.7 14.8 0.905 Not Wasted (weight-for-height z-score ≥ -2) 105.2 15.2     Underweight (weight-for-age z-score < -2) 103 13.2 0.122 Not Underweight (weight-for-age z-score ≥ -2) 105.6 15.5     No moderate acute malnutrition (MUAC ≥ 12.5 cm) 105.7 15.1 0.004* Moderate acute malnutrition (MUAC < 12.5 cm) 99.7 14.5     Vitamin A deficient (retinol binding protein <0.7 µmol/L) 102.1 15.3 0.246 Not vitamin A deficient (retinol binding protein ≥0.7 µmol/L) 105.4 15.2     Iron deficient (serum ferritin < 30 µg/L) 106.2 13.7 0.196 Not iron deficient 104.6 16.2     Iron deficient (serum transferrin receptor > 8.3 mg/L) 104.9 15.3 0.049* Not iron deficient 109.6 12.6     Iron deficient (inflammation adjusted serum ferritin) 102.5 14.5 0.073 Not iron deficient 105.7 15.3     Iron deficient (serum ferritin < 12 µg/L, excluding inflammation) 102 13.7 0.013* Not iron deficient 109.1 12.4 1 Calculated from comparing means using an independent samples t-test * Significant at p < 0.05 4.8. Bivariate Analysis of the Factors Associated with Anaemia Status Table 4.12 presents the bivariate analysis comparing mean values of different continuous variables by anaemia status. Anaemic individuals were found to have statistically significantly (p < 0.05) lower ages and mid-upper arm circumferences. They also had higher counts of malaria parasites in their blood, and higher concentrations of serum ferritin, STfR, CRP and AGP. 64  Table 4.12 Bivariate comparison of anaemia status against continuous participant, household, morbidity, anthropometric and biochemical characteristics in a sample of 6-11 month old infants in Mbala District  Anaemic Not anaemic  Characteristic  Mean SD Mean SD p-value1 Age (months) 8.7 1.8 9.1 1.6 0.001* Age of caregiver (years) 26.3 7.1 27.2 8.3 0.162 Number of individuals in household 5.9 2.4 5.9 2.3 0.963 Number of children under 5 in household 3.4 2.3 3.3 2.1 0.571 Malaria parasites per 200 white blood cells 37.7 287.6 4.6 70 0.037* Start of complementary feeding (months) 5.8 1.1 5.9 1 0.137 Wasting (weight-for-height z-score) -0.1 1.1 0 1 0.591 Stunting (height-for-age z-score) -1.6 1.1 -1.4 1 0.135 Underweight (weight-for-age z-score) -1 1.1 -0.9 1 0.091 Mid-upper arm circumference (cm) 13.7 1.1 13.9 1 0.036* Serum ferritin (µg/L) 71.7 80.9 44.8 40.7 <0.001* Serum transferrin receptor (mg/L) 17.6 8.1 13.4 4.3 <0.001* C-reactive protein (mg/L) 14.1 22.1 6.5 11.8 <0.001* Alpha-1-acid glycoprotein (g/L) 1.6 0.8 1.3 0.5 <0.001* Retinol binding protein (µmol/L) 1.3 0.6 1.3 0.5 0.27 1 Calculated from comparing means using an independent samples t-test * Significant at p < 0.05 Table 4.13 presents the results odds ratios for anaemia against various categorical characteristic of children in the sample. Being female, not having signs of inflammation, having received a vitamin A capsule in the previous 6 months, meeting minimum dietary diversity, the minimum acceptable diet, and having consumed iron-rich foods in the previous 24 hours were all associated with statistically significant (p < 0.05) lower risk of being anaemic, with the largest effect being achieving minimum dietary diversity, with an odds ratio of 0.44 (0.30-0.65). Having had fever in the previous two weeks, having malaria parasitism, being underweight, and moderate acute malnutrition as determined by MUAC were associated with statistically significant (p < 0.05) increases in the odds of being anaemic, with the largest effect being malaria parasitism, with an odds ratio of 4.33 (2.16-8.70). Also of note was the finding that there 65  was no statistically significant change in the odds of being anaemic for individuals who were classified as iron deficient, regardless of the underlying definition, nor was whether or not the child slept under a mosquito net. 66  Table 4.13 Bivariate comparison of anaemia status against categorical participant, household, morbidity, anthropometric and biochemical characteristics in a sample of 6-11 month old infants in Mbala District Characteristic Odds Ratio      (95% CI) Total N Sex (female versus male) 0.71 (0.52-0.97)* 630 Caregiver has attended school 0.95 (0.65-1.39) 631 Household has source of income 0.61 (0.35-1.08) 628 Household has land for cultivation 1.13 (0.48-2.65) 631 Household has garden for fruit and/or vegetables 0.81 (0.59-1.11) 630 Household owns animals or livestock 1.09(0.72-1.64) 626 Household treats water 0.86 (0.62-1.2) 631 Caregiver washes hands with soap or ash 0.88 (0.55-1.4) 630    Child diarrhea in the previous two weeks 1.26 (0.92-1.74) 631 Child cough in the previous two weeks 1.10 (0.78-1.56) 631 Child fever in the previous two weeks 1.88 (1.36-2.59)* 628 Child sleeps under a mosquito net 0.79 (0.56-1.11) 628 Child has malaria parasites 4.33 (2.16-8.70)* 625 No inflammation (AGP ≤ 1 g/L and CRP ≤ 5 mg/L) 0.63  (0.44-0.90)* 625 Received a vitamin A capsule in previous 6 months 0.68 (0.47-0.99)* 628 Received a deworming tablet in previous 6 months 0.85 (0.51-1.42) 629    Child has begun complementary feeding 0.62 (0.25-1.54) 631 Child meeting minimum dietary diversity 0.44 (0.30-0.65)* 631 Child meeting minimum meal frequency 0.92 (0.67-1.26) 628 Child meeting minimum acceptable diet 0.57 (0.38-0.86)* 628 Child has consumed iron rich food previous 24 hours 0.56 (0.40-0.80)* 631 Caregiver has heard of anaemia 0.95 (0.59-1.52) 631 Caregiver has been spoken to about IYCF 1.36 (0.96-0.191) 631    Stunted (height-for-age z-score < -2) 1.32 (0.93-1.87) 629 Wasted (weight-for-height z-score < -2) 1.50 (0.51-4.43) 627 Underweight (weight-for-age z-score < -2) 1.65 (1.05-2.58)* 628 Moderate acute malnutrition (MUAC < 12.5 cm) 2.46 (1.31-4.59)* 631 Vitamin A deficient (RBP <0.7 µmol/L) 1.81 (0.82-4.03) 620 Iron deficient (serum ferritin < 30 µg/L) 0.75 (0.55-1.04) 620 Iron deficient (STfR > 8.3 mg/L) 1.17 (0.637-2.15) 620 Iron deficient (inflammation adjusted serum ferritin) 1.29 (0.80-2.07) 620 Iron deficient (excluding inflammation) 2.23 (0.89-5.61) 620 * Significant at p < 0.05 67  4.9. Linear Modelling of Haemoglobin Concentrations Table 4.14 presents the terms and coefficients of a linear model which was fit to the data. Due to missing terms, 615 children were included in the model. The model, which included age, sex, achieving minimum dietary diversity, MUAC, STfR, CRP, and the presence of malaria parasitism, was statistically significantly, F (7, 607) = 26.16, p < 0.001. The R-square value for the model was 0.232, and the adjusted R-square value was 0.223. Malaria parasitism (B = -4.09), failing to achieve the minimum dietary diversity (B = -3.29), being male (B = -2.22), lower MUAC (B = -1.59), increased STfR concentration (B = -0.71), decreased age in months (B = -0.70), and increased CRP concentration (B = -0.14) were all associated (p < 0.05) with a decrease in haemoglobin concentration. The most negative standardized beta was for STfR concentration at -0.33. Serum ferritin concentration was not included in the model as it was deemed too highly confounded by inflammation; it had a negative beta even when both CRP and AGP were included in the model. Table 4.14 Linear Regression Model of Haemoglobin (Hb) Concentrations (g/L) of Children 6-11 Months in Mbala District, Zambia   95% CI for B   Variable Unstandardized B Lower Upper Standardized B p-value (Constant) 86.28 70.20 102.36 --- --- Age (months) 0.70 0.06 1.33 0.08 0.038 Being female 2.22 0.04 4.40 0.07 0.047 Minimum dietary diversity 3.29 0.66 5.93 0.09 0.017 MUAC (cm) 1.59 0.56 2.62 0.11 0.003 STfR (mg/L) -0.71 -0.87 -0.55 -0.33 <0.001 CRP(mg/L) -0.14 -0.20 -0.08 -0.17 <0.001 Malaria parasitism -4.09 -8.01 -0.18 -0.08 0.048  68  4.9.1. Assumption Checking of the Linear Model As the Durbin-Watson statistic was 1.935, it was accepted that there was no first order correlation between the residuals, or errors, of the model. Furthermore, the Tolerance values ranged from 0.847 to 0.976, suggesting that there is no evidence of collinearity between the independent variables. In checking for extreme points and outliers, there was one case with a standardized residual of 3.01 and one with a standardized residual of -3.182. However, upon closer inspection there was nothing obviously faulty with these data points, and so they were left in the model. The leverage values for model ranged from 0.003 to 0.062, suggesting that there were no leverage points, and the Cook’s Distances ranged from 0.000 to 0.041, suggesting that there were no extreme influence points. Figure 4.1, shown below, suggests that the assumption of homoscedasticity was not met, as the figure shows dispersion, with the residuals increased for larger predicted values. However, given that a heteroscedasticity consistent standard error estimator was used, the significance of the individual terms is still valid.  Figure 4.1 also shows that the assumption of linearity of the model was met, as shown by the best fit line in red. The assumption of linearity was also tested and met for each of the individual independent variables. 69  Figure 4.1 Standardized Predicted Haemoglobin Values versus Standardized Residuals of a Linear Regression Model of Haemoglobin (Hb) Concentrations of Children 6-11 Months in Mbala District, Zambia  Finally, Figure 4.2, below, shows the normal P-P plot of the regression standardized residuals, and based on its shape it was concluded that the residuals are normally distributed. 70  Figure 4.2 Normal P-P Plot of Regression Standardized Residuals of a Linear Regression Model of Haemoglobin (Hb) Concentrations of Children 6-11 Months in Mbala District, Zambia   4.10. Logistic Modelling of Anaemia Status Table 4.15 presents the terms and coefficients of a logistic model, which was fit to the data. Due to missing data, 612 children were included in the model. The model, which included age, sex, log parasite density, minimum dietary diversity, STfR concentration, CRP concentration, fever, and IYCF instruction was statistically significant, χ2(8) = 122.597, p < 71  .0001. The Nagelkerke R-square for the model was 0.244, the sensitivity of the model was 75.0% and the specificity was 59.2%. All terms were statistically significant at p < 0.05, except for sex, which was forced into the model. Increased age (OR = 0.84)) and achieving minimum dietary diversity (OR = 0.58) statistically significantly decreased the risk of being anaemic, whereas fever in the previous 2 weeks (OR = 1.64), the caregiver having been counselled by a healthcare worker about IYCF practices (OR = 1.61), increased log malaria parasite density (OR =1.42), increased STfR concentration (OR = 1.10), and increased CRP concentration (OR = 1.02) all statistically significantly increased the risk of being anaemic. Fever in the previous 2 weeks had the highest odds ratio and minimum dietary diversity had the lowest odds ratio. Table 4.15 Logistic Regression Model of Anaemia Status of Children 6-11 Months in Mbala District, Zambia      95% CI for OR Variable B Wald p-value OR Lower Upper Constant 0.08 0.02 0.893 1.08 --- --- Age (months) -0.18 10.13 0.001 0.84 0.75 0.93 Being female -0.18 1.03 0.311 0.83 0.58 1.19 Log10 (malaria parasites / 200 WBC) 0.35 4.19 0.041 1.42 1.01 1.98 Minimum dietary diversity -0.55 6.37 0.012 0.58 0.38 0.88 STfR (mg/L) 0.10 32.21 <0.001 1.10 1.07 1.14 CRP(mg/L) 0.02 9.90 0.002 1.02 1.01 1.03 Fever in last 2 weeks 0.50 7.36 0.007 1.64 1.15 2.35 Caregiver counselled on IYCF 0.48 5.80 0.016 1.61 1.09 2.38  4.10.1. Assumption Checking of the Logistic Model The Hosmer and Lemeshow goodness of fit test was not statistically significant (p = 0.508) and so there was a failure to reject the null hypothesis that the model was a good fit for the data. The Box-Tidwell procedure with a Bonferroni correction was used to test the assumption of linearity between the continuous independent variable and the logit of the 72  dependent variable. From this test, none of the interaction terms between the continuous independent variables and their log transformation were statistically significant, and so it was concluded that the assumption of linearity was met. 73  5. Discussion The following chapter will discuss the results of the thesis and put them into the context of previous findings, starting with the findings related to anaemia and iron deficiency. It will discuss general participant characteristics including morbidity, and IYCF practices, as these are key in understanding some of the results for the modelling, as well as the potential issue of generalizability. It will then conclude with a discussion of some of the limitations of the research, ending with a brief section on areas for future research. 5.1. Prevalence of Anaemia One of the major findings of the thesis was the high prevalence of anaemia, with 57% of the sample testing positive for mild or moderate anaemia based on concentrations of haemoglobin. This is well above the 40% prevalence cut-off used by UNICEF to indicate a “severe” health crisis (8), although it is below the national prevalence rate for  anaemia among children under 12 months, which according to the 2012 Malaria Indicator Survey was 70% , although this prevalence also include severe anaemia (7). There was a higher prevalence of moderate anaemia, which was found in 33% of the children, as opposed to mild anaemia, which was found in 25% of the children. The fact that moderate anaemia was found to be more prevalent than mild anaemia has implications for the health of the children, as it indicates a higher disease burden.  As is discussed further in limitations, the fact that children with severe anaemia were excluded from the study means that the 57% prevalence rate is likely an underestimate of the true population prevalence of anaemia, although it is difficult to say with any certainty to what extent this would be the case. It is possible to arrive at an estimate of severe anaemia of ~3% based on field notes from one of the Zambian field editors tracking health exclusion, but this is an 74  approximation only and should be considered only with extreme caution. The most recent Malaria Indicator Survey from 2012 found a 7% prevalence of severe anaemia in children under 5, but this was calculated using an elevated severe anaemia cut-off of 80 g/L Hb (7), and so is essentially an overestimate. Unfortunately, this cut-off is used for all of the Malaria Indicator Surveys used in Zambia. A separate study focused on vitamin A supplementation in Zambia, using the standard cut-off and a nationally representative sample, found a prevalence of severe anaemia of only 2%, but as this study is from 2003 its usefulness is limited (9).  5.2. Prevalence of Iron Deficiency and Iron Deficiency Anaemia Another key finding was that the prevalence of iron deficiency, and thus iron deficiency anaemia, in the population was highly dependent on the assessment method that was used, a finding which has been reported elsewhere. The lowest prevalence of iron deficiency, 14%, was found when using adjusted serum ferritin cut-offs from Thurnham and McCabe in an attempt to account for the high rate of inflammation in the population, as discussed previously (52). This is critical because while serum ferritin has excellent specificity, its sensitivity is compromised during inflammation as a result of it being a positive acute phase responder (52). The 14% prevalence figure is roughly similar to the 13% prevalence rate which is obtained when using the standard cut-off of serum ferritin < 12 µg/L, while only analyzing those children who are not exhibiting inflammation. This number must itself be treated with caution, as it is a result of excluding 74% of the sample, and is by its nature very biased, as it is possible that individuals that are sick may have higher rates of iron deficiency than the broader population. A third method of using serum ferritin to assess iron deficiency, which makes use of a cut-off of < 30 µg/L, resulted in a notably higher prevalence of iron deficiency of 42%. 75  The highest prevalence for iron deficiency was found when using STfR. Based on STfR > 8.3 mg/L, 93% of the children are iron deficient, which stands in marked contrast to the findings when using serum ferritin concentrations. This disparity has been reported in other settings, and represents an unresolved issue in the diagnosis of iron deficiency in the context of high inflammation (24,88). However, the results from STfR must also be interpreted with caution; while STfR is currently not considered an acute phase responder (102), it is possibly increased by malaria parasitism (91) and α-thalassaemia (94).  It is therefore difficult to say with certainty what the “true” prevalence of iron deficiency is in the population, especially given that previous attempts to assess iron deficiency in Zambian children have yielded inconsistent results, that are potentially very location specific (14,76–79). Certainly, the high specificity of the traditional serum ferritin cut-off suggests that 13% is a reasonable lower limit, and the potentially high prevalence from STfR indicates the need for follow-up studies, as will be discussed below. One possibility may be to go with the elevated serum ferritin cut-off of < 30 µg/L, as recommended by the WHO (99), although this cut-off is itself problematic as it lacks a strong theoretical backing. It should also be noted that all the findings from serum ferritin are in stark contrast to the estimate, reported by the WHO, that in a given population the rate of iron deficiency can be expected to be approximately 2.5 times the prevalence of anaemia (23).  The uncertainty in iron deficiency is reflected in the prevalence of IDA, which is either 8% (Thurnham and McCabes serum ferritin cut-offs), 9% (when excluding inflammation), 22% (serum ferritin < 30 µg/L) or 53% (STfR > 8.3 mg/L). For the lowest prevalence this would mean that 14% of the anaemia in the population could be characterized as being IDA, and for the highest prevalence this implies that 93% of the anaemia in the population is due to IDA, though 76  other factors could also be contributing. Thus it is difficult to draw firm conclusions, other than the need for further research to more accurately define the rate of IDA in the population as well as determine how best to assess iron deficiency in similar contexts. This information is critical as it is relevant to the design of any future interventions seeking to reduce the rate of anaemia in the region. It is also critical to better understand the potential health burdens associated with anaemia in the region, as IDA is characterized by permanently impaired cognitive, psychosocial and psychomotor skill development when found in this age group (11,70,73)  In, the absence of clear guidelines on how best to adjust for inflammation, one solution, although time consuming, would be to measure a sample of anaemic children from the region and split them into two groups: one receiving a placebo and one receiving an iron supplement (49). These groups could have their haemoglobin concentrations measured again in several months, and based on the relative changes in haemoglobin concentrations one could determine the rate of iron deficiency (49). Based on a similar study design, iron deficiency was ruled out as a major cause of anaemia in the Macha region of Zambia (78). Another option would involve determining whether there is a strong seasonal component to inflammation rates in the region; if there was it would be advisable to measure serum ferritin during the period of time when inflammation was lowest, thus reducing its confounding influence (103). However, given that the current study was already undertaken outside of the peak malaria transmission period, running from December through May (104), it is  not immediately obvious when a better time might be. 5.3. Participant and Household Characteristics Before turning specifically to the results of the bivariate analysis and the modelling, it is important to discuss some of the broader results concerning the participants and their households, 77  as these are important covariates and relate to the broader issue of generalizability. One finding of note was the high burden of disease in population, as seen from caregiver reports and biochemical analysis. At 56% for diarrhea, 72% for cough, and 56% for fever in the previous two weeks, there was a significant burden of disease. This is in stark contrast to findings from both the Malaria Indicator Survey from 2012, which found a 37% prevalence of fever for children under 5 in the Northern Province and a 24% prevalence nationally (7). It is also in contrast to the 2007 DHS, which found an 18% prevalence rate for diarrhea among children under 5 in the Northern Province and a 16% prevalence nationally (4).  The high prevalence of morbidity was further reflected in the testing of CRP and AGP concentrations, which indicated a 74% prevalence of inflammation. Taken as a whole these findings suggest caution in generalizing the findings from this research to other children in Zambia, as the children in the sample may be sicker than is typical. It is also unclear whether these findings are representative of the typical state of health of the local population, although the high concentrations of AGP in the population suggest a chronic disease burden, rather than a purely transitory phenomenon (52). This high rate of infection and thus inflammation is also behind the difficulty in establishing a “true” prevalence of iron deficiency or IDA as discussed above. However, the malaria burden in the population, at 10% parasitism matches the findings from the Malaria Indicator Survey from 2012, which found a 10% rate of parasitism nationally in children under 12 months (7). The proportion of children under 12 months sleeping under a mosquito net the previous night, at 69%, was also the same in both the present study and the Malaria Indicator Survey (7). 78   The dietary quality of the children in the sample also seems to be poorer than that found nationally, or even at the provincial level, based on the proportions of children achieving the minimum dietary diversity and the minimum acceptable diet, which are 37% and 25% nationally, and 22% and 18% in the study sample (4). However, the proportion of children achieving the minimum meal frequency was higher, at 56%,  than the national rate of 49% (44). Furthermore, the prevalence of vitamin A deficiency, as determined via concentrations of RBP was very low, at only 5%, perhaps reflecting the relatively high coverage for vitamin A supplementation, at 74% over the preceding 6 months. Turning to anthropometry, the prevalence of stunting, at 30%, was nearly identical to findings at the national level, in which 26% of children 6-8 months and 33% of children 9-12 months are stunted, although the study sample does seem to have a lower prevalence of severe stunting (4). The finding that boys are more stunted than girls has also been found nationally (4) and is in fact the case across much of English Speaking sub-Saharan Africa for reasons which are still unclear (105). There was a much lower prevalence of moderate wasting in the sample, at 2%, than found nationally, where it was found that 8% of children 6-8 months and 12% of children 9-12 months were moderately wasted (4). Due to the exclusion of individuals with severe wasting from the study, it is not possible to comment on how it may compare to national findings. Finally, the study sample was slightly more underweight than has been found nationally. While 16% of children in the study were underweight, only 10% of children 6-8 months and 15% of children 9-12 months were underweight nationally in the previous DHS (4). 5.4. Factors Associated with Haemoglobin Concentration and Anaemia Status The study identified several factors associated with haemoglobin concentration and anaemia status among the study participants. In terms of general participant characteristics, both 79  age and sex were found to effect haemoglobin concentrations, as well as anaemia status, with higher concentrations of haemoglobin found as children aged, and in girls, who had an odds ratio for anaemia of 0.71 when compared to boys. The finding that age is related to haemoglobin concentration and anaemia status is in line with the most recent Malaria Indicator Survey which found a decreasing prevalence of anaemia as age increased (7) as well as in studies elsewhere (81,106,107). The relationship between anaemia and sex has also been found elsewhere (108). There was no relationship between educational status of the primary caregiver and either haemoglobin concentration or anaemia status of the child, in line with previous findings from Zambia (9) and Nigeria (61), but not Tanzania (62). Along a similar vein there was no relationship between whether or not the primary caregiver had heard of either anaemia or iron deficiency and the risk of their child being anaemic, in contrast to a study from Indonesia (48).   It was also found that whether or not a household treated their water had a statistically significant effect on the haemoglobin concentration a child, with a mean haemoglobin concentration that was 3 g/L higher in households that treated their water; however, when analyzing anaemia status, water treatment was no longer statistically significant. The effect of water treatment on haemoglobin status has been seen previously (109) is possibly related to the numerous associations found between haemoglobin concentrations and various forms of morbidity (43) as well as the potential association between water quality and diarrhea (110). Notably, diarrhea in the previous two weeks was associated with decreased haemoglobin concentrations, paralleling results in Timor-Leste (108), Kenya (111), and Brazil (109), although the relationship was no longer statistically significant when looking at anaemia status. One possible explanation for the effect of diarrhea on haemoglobin is the increased loss of iron due to loss of cells in the intestinal lining during episodes of diarrhea (112).  80  Fever was also associated with lower concentrations of haemoglobin and a higher risk of anaemia, with one possible pathway being malaria. Higher parasite loads were associated with reductions in haemoglobin concentrations, and a greatly increased risk of being anaemic (OR = 4.33). This parallels findings from other Zambian studies which also found an association between malaria and anaemia (9,74,75,78). Despite this, there was no association between haemoglobin concentrations or anaemia status and whether or not a child had slept under a mosquito, in line with a study from the Luangwa District of Zambia (75).  Interestingly, there was no association between either anaemia status or haemoglobin concentration and whether or not a child had received de-worming tablets in the previous 6 months, which contradicts earlier findings from Zambia (9). The relationship between infection, and more broadly inflammation, on haemoglobin concentrations and anaemia status was particularly strong. Both CRP and AGP concentrations were negatively correlated with haemoglobin concentrations, as has been found previously (81), although not in all cases (113). This relationship was also reflected in the finding that anaemic individuals had higher concentrations of both chemicals, which was particularly the case for CRP. Combining both chemicals into a general indicator of inflammation, it was found that those who were experiencing inflammation manifested by either elevated CRP or AGP had a higher risk of being anaemic than those who were not (OR = 1.59). Several factors could be at play in these relationships. In addition to possibly reflecting a number of morbidities, such as malaria and other forms of parasitism, as well as the conditions which are causing fever and diarrhea, it is possible that these findings indicate the effect of ACI on the population (43). This would suggest that one way to improve the anaemia status of the population would be to examine interventions that reduce the general burden of disease. 81  The importance of dietary practices was seen in the finding that children achieving the minimum dietary diversity had a lower risk of being anaemic than those that were not, as has been found elsewhere (114). This could reflect the important of various micronutrient deficiencies, as dietary diversity has been found to be a good predictor of the micronutrient density in a number of contexts (45) and a large number of micronutrient deficiencies have been linked to anaemia (103). The importance of dietary diversity was strong enough to drive a statistically significant relationship between the IYCF indicator minimum acceptable diet and anaemia, despite minimum meal frequency, which is a component of minimum acceptable diet, not being statistically significant in contrast to findings in rural China  (114).  Potentially related to dietary quality was the finding that while higher values for MUAC, the stunting z-score and the underweight z-score were associated with increased haemoglobin concentrations, the wasting z-score was not. This relationship was notable for MUAC: children with MUAC above 12.5cm had an average haemoglobin concentration of 106 g/L whereas children with MUAC between 11.5cm and 12.5cm, indicative of moderate acute malnutrition, had average haemoglobin concentrations of only 100 g/L. The risk of being anaemic was 2.5 times greater for children with moderate acute malnutrition than those without. Given this, the lack of a relationship between wasting and anaemia is difficult to explain, as MUAC is generally considered a proxy for wasting. Previous research has found an association between underweight and anaemia in South African infants 6-11 months (115), and the relationship between stunting and anaemia corresponds to findings in that same South African cohort (115) as well as ones in Nepal (107) and Kenya (111). Finally, there was a strong negative relationship between haemoglobin concentration and both serum ferritin concentrations and STfR concentrations. While this is the expected result in 82  the case of STfR (103), the negative correlation coefficient for serum ferritin is noteworthy in that it demonstrates how confounded this iron indicator was by the acute phase process (81). In a healthy population one would expect there to be a positive correlation between serum ferritin and haemoglobin, as higher concentrations of the former would imply adequate iron reserves in the body (23). However, in this sample anaemic individuals actually had higher serum ferritin than those without anaemia.  Furthermore, the negative association between serum ferritin concentrations and haemoglobin concentration remained even after adjusting for the effect of inflammation using the method suggested by Thurnham and McCabe (52). As a result one must conclude that serum ferritin is likely of limited use in attempting to diagnose iron deficiency in this context. The negative correlation between STfR concentration and haemoglobin concentration likely represents the effect of worsening iron status in the face of depleted iron stores (103), although to a certain extent it is possible that this is also being caused by the presence of α-thalassaemia (94), complicating its interpretation. Despite this, STfR is still likely the preferred indicator when compared to serum ferritin in the present setting as it is generally viewed as less effected by inflammation (103). While one study from Lao PDR did find that individuals had elevated STfR concentrations during the convalescence phase of an infection, they hypothesised that this may have been the result of increased erythropoiesis, rather than merely an artifact of the inflammation (116). The difficulty in choosing cut-offs is demonstrated by the finding that there was no statistically significant difference between the risk of being anaemic for iron deficient and non-iron deficient individuals, regardless of the definition of iron deficiency used, although individuals who were diagnosed as iron deficient according to their concentration of STfR did have a statistically significant reduction in their haemoglobin concentrations. 83  5.5. Modelling of Haemoglobin Concentration and Anaemia Status  The modelling of haemoglobin concentration and anaemia status are particularly valuable given that there are potentially a large number of covariates and confounding variables that explain haemoglobin concentrations and anaemia, although the linear model of haemoglobin concentration and the logistic model of anaemia status tell slightly different stories.  In the linear model both age and sex are statistically significant, paralleling the findings from bivariate analysis and other regions (81,108,117), although the effect size of being female was reduced to only a 2 g/L increase in haemoglobin concentrations. This compares to a 1 g/L  effect size seen in Timor-Leste (108). Likewise the effect size for age was limited at less than 1 g/L per month increase. The importance of dietary factors was borne out by the statistically significant terms of achieving dietary diversity and MUAC, with a 1 centimeter increase in MUAC resulting in an approximately 2 g/L increase in haemoglobin. This term had a standardized beta of 0.11, which was greater than that for age, sex, and achieving minimum dietary diversity. Furthermore, each 1 mg/L increase in STfR concentrations resulted in a nearly 1 g/L reduction in haemoglobin, an effect seen elsewhere (81) and in line with expectations (103). Given that the mean concentration of STfR is 16 mg/L and that it ranges from 6 to 40 mg/L, this represents a major contributor to the concentration of haemoglobin in the study sample. This is illustrated by the standardized beta of -0.33 for STfR, which is greater than any other independent variables, and suggests that iron status has a large, important, effect on haemoglobin in this population, although this could also be responding to the effect of α-thalassaemia (94). The importance of infection and morbidity is indicated by the statistically significant terms for CRP concentration and malaria parasitism, the former of which had the second largest 84  standardized beta of any term in the model, at -0.17. The importance of malaria in the modelling of anaemia has been reported in numerous other African contexts (61,106,111).  Several terms which were significant during bivariate analysis were not included in the linear model as they were no longer statistically significant, these included diarrhea, fever, and water treatment. The finding that they were no longer significant implies that their effect on haemoglobin concentration was explained through the other variables in the model, possibly CRP concentration and malaria parasitism. However, the model only explained approximately 22% of the total variation in haemoglobin concentrations, suggesting that a number of important covariates were not included. Another concern was the presence of dispersion when analyzing the relationship between the standardized predicted values of the model and the standardized residuals. At higher predicted models, that was an increase in the absolute value of the residuals, indicating that the model was better suited to explaining haemoglobin values at the lower end of their range. The logistic regression model of anaemia status was similar to the linear model, but included some different terms. MUAC was dropped, but fever and whether a caregiver had been spoken to by a healthcare worker about IYCF were included, and the malaria term was modified to reflect the log of the number of malaria parasites per 200 white blood cells. Sex was forced into the model despite not being statistically significant. Fever has been found to be an important predictor of anaemia status in Nigeria (61) and Kenya (111). Given that the sensitivity of the model was 75.0% and the specificity was 59.2%, the model was fairly good at correctly identifying anaemia individuals, but suffered by a somewhat high rate of false positives.  From the exponentiated betas there was a 10% increase in the risk of being anaemic for each 1 mg/L increase in STfR, which alongside previous findings supports the contention that 85  iron deficiency is an important factor in the local context.  Most surprising, and the more difficult to explain, was the result that being spoken to about IYCF, specifically breastfeeding and how to feed a child, by a healthcare worker resulted in an increase in the risk of being anaemic, with an odds ratio of 1.61. This is counterintuitive, as one might expect that such messaging would lower the risk of anaemia, rather than increase it, especially as there are numerous findings associated improved IYCF practices with child health (46,64,114).  One likely explanation is that the variable is actually highly correlated with a third, unspecified, variable which is driving the relationship, such as the number of contacts with the health system, or that children which are in some way unwell are more likely to attract the attention of healthcare workers, resulting in a higher rate of messaging being directed at these individuals. Unfortunately there is no way to be certain given the present study design. The other terms, CRP concentration, achieving minimum dietary diversity and the log of the parasites all behaved similarly to their counterparts in the linear model, with the largest effect size for minimum dietary diversity, with children achieving this standard having a risk 0.58 times lower than those who did not. This further implies that diet is a major driver of anaemia, in line with other findings (50,107,114). Overall the models emphasize the importance of inflammation, malaria, dietary quality, age, and iron status in understanding anaemia  5.6. Limitations There are a number of limitations to the study design. Roughly speaking these limitations can be divided into those involving sampling and those involving potential confounders and explanatory variables which were not measured. One of the largest limitations is that the study made use of a convenience sample, as opposed to a random sample. The decision to do so was based on feedback from local partners who stated that, given the amount of suspicion in some 86  rural areas towards health projects generally, and those involving blood draws specifically, it would not have been feasible to select individuals from village lists, as it would result in low participation. This raises the issue that the sample may not be representative of the population in the selected Catchment Areas. However, based on reports from local partners at the District Medical Office, there were only 3973 eligible children in the selected zones. Given that 631 were in the end recruited, it means that approximately 16% of the eligible population was recruited, which represents a sizable portion, and somewhat alleviates concerns around the lack of random sampling. Of course, there remains the possibility that the children of parents who did not want to participate are in some ways different from the children of parents who did, particularly in terms of their interactions with the health system. Another issue with the sampling relates to the exclusion criteria. Given that children with severe acute malnutrition were excluded from the dataset, it is impossible to tell to what extent the level of malnutrition seen in the sample is actually reflective of the population. More critically, the study excluded children with severe anaemia. This reduces the scope of the research, preventing any discussion of the factors associated with severe anaemia in this population. However, this is potentially beneficial, as it allows for a more focused analysis: the causes of mild and moderate anaemia are not necessarily the same as the causes of severe anaemia, as seen when comparing severe malarial anaemia (53) and the mild anaemia brought on by some haemoglobinopathies, such as α-thalassaemia (59). Another key consideration is that this exclusion means that the prevalence of anemia found in the sample may be an underestimate when compared to the wider population.  There are also a number of potential confounding variables and covariates which unfortunately were not measured due to limited resources, some more critical than others. This is 87  reflected in the R-square value for the linear regression model, which was only 0.223, meaning that a large amount of variation in haemoglobin concentrations remains unexplained. It would have been particularly useful to have more biochemical measures of micronutrient status, particularly for B12, folate, riboflavin, and to a lesser extent copper, and zinc as deficiencies in any of these may cause anaemia (8). Another unmeasured variable, which may have been an important explanatory variable, as well as a confounder of STfR, is the presence of α-thalassaemia and/or the sickle cell trait. These conditions, which have been found in other Zambian studies (58–60), can result in elevated STfR levels (93–95), and so reduces its specificity. This could be a possibly explanation for the marked difference in iron deficiency when using STfR as opposed to serum ferritin.  Mean corpuscular volume, or MCV, would also have improved the interpretation of iron status, as it would have indicated whether or not any anaemia was microcytic, as is seen in IDA. It would also have been ideal to measure the presence of hookworms, Schistosomiasis, and other parasites (8), as well as for the presence of maternal anaemia (64). Of course, a final limitation is a result of the overall study design. Given that the study is a cohort study there is no way to establish causal pathways, just correlations. 5.7. Future Research The findings and the limitations of the study suggest a number of follow-up studies. Some are broad in scope, such as a need to investigate the best way to define iron deficiency in the context of high inflammation and malaria. In the absence of such information, a randomized control trial with iron supplementation, measuring haemoglobin concentration before and after, could be undertaken to more definitively determine the prevalence of iron deficiency in the area. Ideally such a study would also measure CRP, AGP, STfR and serum ferritin, along with MCV. 88  It would also be useful to more accurately determine the prevalence of α-thalassaemia and the sickle cell trait in the area, as well as the prevalence of various parasites such as hookworms, Ascaris, and Schistosomiasis, and the prevalence of other key deficiencies, such as B12, B6, riboflavin, and folate, as this information is currently unavailable. These variables are important to understanding the etiology of anaemia, but there is currently very little information on them in Zambia. Such information could be used to guide studies on these variables in other regions, as well as being useful in health interventions. Also important is the need to understand the relationship between IYCF messaging and anaemia status, as the present findings were counterintuitive at first glance. If this variable was merely reflecting another variable, as is suspected, it is important to understand what it may have been. Not only will this information be useful for understanding anaemia, it will also be valuable when measuring IYCF practices in the future. Another study could further examine the effect of water treatment, and sanitation more generally on anaemia. Results from such a study could be particularly useful for designing novel interventions which target non-nutritional causes of anaemia.  Finally, it is also important properly understand the etiology of anaemia in the country as a whole. While this is currently done to a certain extent in the Malaria Indicator Surveys, anaemia is not their focus. Such a study would incorporate random sampling, and measure the variables mentioned above.   89  6. Conclusion This study set out to determine the prevalence of anaemia, iron deficiency, and IDA among children 6-11 months in Mbala District, Northern Province, Zambia, as well as to explore the factors associated with mild and moderate anaemia. Based on concentrations of haemoglobin < 110 g/L, 57% of the population was anaemic. However, attempts to determine the prevalence of iron deficiency were complicated by the high rates of inflammation in the population: using a definition of CRP > 5 mg/L and/or AGP > 1 g/L, 74% of the population exhibited inflammation.  The effect of this inflammation is seen in the disparate rates for iron deficiency, ranging from 13-93%, and IDA, ranging from 8-53% depending on the definition used, as well as the finding that there was a negative correlation between serum ferritin and haemoglobin concentrations, the opposite of what would be expected from theory. In the end, iron status, regardless of cut-off, was not a statistically significant correlate of anaemia status, although this is more probably a result of issues with the cut-offs in this context, rather than an indication that iron deficiency and anaemia are not linked in the population. Logistic regression found that increased age in months (OR = 0.84) and achieving minimum dietary diversity (OR = 0.58) were associated with a reduced risk of being mildly or moderately anaemic, while increased log malaria parasite loads (OR = 1.42), STfR concentrations (OR = 1.10), CRP concentrations (OR = 1.02), fever in the previous 2 weeks (1.64), and having been counselled on IYCF by a health care professional (OR = 1.61) were associated with an increased risk of being mildly or moderately anaemic. These findings are consistent with the current understanding of the etiology of anaemia, with the exception of the relationship between IYCF counselling and anaemia status, which is difficult to explain, and possibly the result of an unidentified mediating factor. There was no relationship between 90  whether or not a child slept under a mosquito net and anaemia status, suggesting that an intervention targeting malaria to reduce anaemia may have to focus on a different mechanism. Furthermore, children that lived in households that treated their water had on average 3 g/L higher concentrations of haemoglobin than those who did not (p = 0.021), and children with diarrhea in the previous two weeks had on average 3 g/L lower haemoglobin concentrations (p = 0.042), although these relationships disappeared in multivariate analysis. There were some limitations to study design which may have affected the results: the lack of randomization, the exclusion of children with severe anaemia, and the fact that certain covariates, such as haemoglobinoptahies, were not measured. However, the findings suggest the importance of diet, poor iron status, malaria, and infection in understanding the etiology of anaemia in this region of Zambia. Less clear, but still worthy of further research, is the potential relationship between anaemia and sanitation suggested by the bivariate findings related to water treatment and diarrhea. This information is critical as there is currently only limited research on the specific etiology of anaemia in the Zambian setting. It can be used to implement future interventions as well as to guide the design of studies intended to establish causal pathways.  Overall, the high prevalence of anaemia stresses the importance of continued efforts to understand its causes in detail as well as the importance of programs aimed at tackling it. Possible interventions could focus on sanitation and water quality, improving the quality of the diet through diversification, fortification and education, and reducing the prevalence of diseases such as malaria.  91  Bibliography 1.  NFNC. The First 1000 Most Critical Days Three Year Programma 2013-2015. Lusaka, Zambia; 2012.  2.  UNSCN. State of the SUN Movement Progress Report (September 2013). Geneva, Switzerland; 2013.  3.  UNICEF. The State of the Word’s Children 2009: Maternal and Newborn Health. New York, NY; 2008.  4.  CSO. The 2007 Zambia Demographic and Health Survey. Lusaka, Zambia; 2009.  5.  UNICEF. Tracking Progress on Child and Maternal Nutrition. New York, NY; 2009.  6.  World Bank. Nutrition at a Glance: Zambia.  7.  Ministry of Health. Zambia National Malaria Indicator Survey 2012. Lusaka, Zambia; 2013.  8.  WHO. Worldwide prevalence of anaemia 1993-2005. Geneva, Switzerland; 2008.  9.  Clewes C, Kankasa C. Report of the National Survey to Evaluate the Impact of Vitamin A Interventions in Zambia In July and November 2003. 2003.  10.  Dewey K, Begum K. Long term consequences of stunting in early life. Matern Child Nutr. 2011;7:5–18.  11.  Horton S, Ross J. The economics of iron deficiency. Food Policy. 2003;28:51–75.  12.  Ouédraogo HZ, Zeba A, Dramaix-Wilmet M, Donnen P. Moderate-to-severe anaemia due to afebrile Plasmodium falciparum infection in children aged 6-23 months from the rural district of Kongoussi, Burkina Faso. J Trop Pediatr. 2008 Dec;54(6):395–400.  13.  Righetti A a, Adiossan LG, Ouattara M, Glinz D, Hurrell RF, N’Goran EK, et al. Dynamics of anemia in relation to parasitic infections, micronutrient status, and increasing age in South-Central Côte d’Ivoire. J Infect Dis. 2013 May 15;207(10):1604–15.  14.  Van Rheenen PF, de Moor LTT, Eschbach S, Brabin BJ. A cohort study of haemoglobin and zinc protoporphyrin levels in term Zambian infants: effects of iron stores at birth, complementary food and placental malaria. Eur J Clin Nutr. 2008 Dec;62(12):1379–87.  15.  PAHO/WHO. Guiding Principles for Complementary Feeding of the Breastfed Child. Geneva, Switzerland; 2003.  92  16.  Gitau R, Makasa M, Kasonka L, Sinkala M, Chintu C, Tomkins A, et al. Maternal micronutrient status and decreased growth of Zambian infants born during and after the maize price increases resulting from the southern African drought of 2001–2002. Public Health Nutr. 2005 Jan 2;8(7):837–43.  17.  National AIDS Council. Zambia Country Report: Monitoring the Declaration of Commitment on HIV and AIDS and the Universal Access. Lusaka, Zambia; 2012.  18.  Central Statistical Office. 2010 Census of Population and Housing National Analytical Report. Lusaka, Zambia; 2012.  19.  WHO. Disease Surveillance for Malaria Control: An Operational Manual. Geneva, Switzerland; 2012 Jan.  20.  Mwape KE, Phiri IK, Praet N, Muma JB, Zulu G, Van den Bossche P, et al. Taenia solium Infections in a rural area of Eastern Zambia-a community based study. PLoS Negl Trop Dis. 2012 Jan;6(3):e1594.  21.  Hautvast JL, Tolboom JJ, Kafwembe EM, Musonda RM, Mwanakasale V, van Staveren W a, et al. Severe linear growth retardation in rural Zambian children: the influence of biological variables. Am J Clin Nutr. 2000 Feb;71(2):550–9.  22.  Chipeta J, Mwansa J, Kachimba JS. Schistosomiasis Disease Burden in Zambian Children: Time for affirmative action is now. Med J Zambia. 2009;36(1):1–5.  23.  WHO. Iron deficiency anaemia: Assessment, Prevention, and Control, A guide for programme managers. Geneva, Switzerland; 2001 Mar.  24.  Engle-Stone R, Nankap M, Ndjebayi A, Erhardt JG, Brown K. Plasma Ferritin and Soluble Transferrin Receptor Concentrations and Body Iron Stores Identify Similar Risk Factors for Iron Deficiency but Result in Different Estimates. J Nutr. 2013;(8):369–77.  25.  Dewey K, Baldiviez L. Safety of Universal Provision of Iron through Home Fortification of Complementary Foods in Malaria-Endemic Areas. Adv Nutr. 2012;3:555–9.  26.  Suchdev P, Ruth L, Woodruff A, Mbakaya C, Mandava U, Flores-Ayala R. Selling Sprinkles micronutrient powder reduces anemia, iron deficiency, and vitamin A deficiency in young children in Western Kenya: a cluster-randomized controlled trial. Am J Clin Nutr. 2012;95:1223–30.  27.  Rah JH, Kraemer K, Steiger G, Bloem MW, Spiegel P, Wilkinson C, et al. Program Experience with Micronutrient Powders and Current Evidence. J Nutr. 2012;142:191S–196S.  28.  WHO. Guideline: Use of multiple micronutrient powders for home fortification of foods consumed by infants and children 6 – 23 months of age. Geneva, Switzerland; 2011.  93  29.  WHO. Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity. Geneva, Switzerland; 2011.  30.  Nestel P. Adjusting Hemoglobin Values in Program Surveys. Washington, D.C.; 2002.  31.  Gibson R. Principles of nutritional assessment. 2nd ed. New York, NY: Oxford University Press; 2005.  32.  Stoltzfus R, Dreyfuss M. Guidelines for the use of iron supplements to prevent and treat iron deficiency anemia. Washington, D.C.: International Life Sciences Institute Press; 1998.  33.  Domellöf M, Dewey K, Lonnerdal B, Cohen R, Hernell O. The diagnostic criteria for iron deficiency in infants should be reevaluated. J Nutr. 2002;132:3680–6.  34.  Emond A, Hawkins N, Pennock C, Golding J. Haemoglobin and ferritin concentrations in infants at 8 months of age. Arch Dis Child. 1996 Jan;74(1):36–9.  35.  Thorsdottir I, Gunnarsson BS, Atladottir H, Michaelsen KF, Palsson G. Iron status at 12 months of age -- effects of body size, growth and diet in a population with high birth weight. Eur J Clin Nutr. 2003 Apr;57(4):505–13.  36.  Johnson-Spear A, Yip R. Hemoglobin difference with comparable iron anemia between black status: justification and white women for race-specific anemia critera. Am J Clin Nutr. 1994;60:117–21.  37.  Khusun H, Yip R, Schultink W, Dillon DHS. World Health Organization Hemoglobin Cut-Off Points for the Detection of Anemia Are Valid for an Indonesian Population. J Nutr. 1999;129:1669–74.  38.  Krebs N, Mazariegos M. Randomized controlled trial of meat compared with multimicronutrient-fortified cereal in infants and toddlers with high stunting rates in diverse settings. Am J Clin Nutr. 2012;96:840–7.  39.  Ministry of Health. Zambia National Malaria Indicator Survey 2010. Lusaka, Zambia; 2010.  40.  Earl R, Woteki CE. Iron Deficiency Anemia: Recommended Guidlines for the Prevention, Detection, and Management Among U.S. Children and Women of Childbearing Age. Earl R, Woteiki CE, editors. Washington, D.C.: National Academy Press; 1993.  41.  Ministry of Health. Zambia National Malaria Indicator Survey 2006. Lusaka, Zambia; 2009.  42.  Ministry of Health. Zambia National Malaria Indicator Survey 2008. Lusaka, Zambia; 2009.  94  43.  Thurnham DI, Northrop-clewes CA. Infection and the etiology of anemia. In: Kraemer K, Zimmermann MB, editors. Nutritional Anemia. Basel, Switzerland: Sight and Life Press; 2007. p. 231–56.  44.  WHO. Indicators for assessing infant and young child feeding practices. Part 3: Country Profiles. Geneva, Switzerland; 2010.  45.  Moursi MM, Arimond M, Dewey KG, Tre S, Ruel MT, Delpeuch F. Dietary Diversity Is a Good Predictor of the Micronutrient Density of the Diet of 6- to 23-Month-Old Children in Madagascar. J Nutr. 2008;138:2448–53.  46.  Arimond M, Ruel MT. Dietary Diversity Is Associated with Child Nutritional Status : Evidence from 11 Demographic and Health Surveys. J Nutr. 2004;134:2579–85.  47.  Stevens G a, Finucane MM, De-Regil LM, Paciorek CJ, Flaxman SR, Branca F, et al. Global, regional, and national trends in haemoglobin concentration and prevalence of total and severe anaemia in children and pregnant and non-pregnant women for 1995–2011: a systematic analysis of population-representative data. Lancet Glob Heal. 2013 Jul;1(1):e16–e25.  48.  Souganidis ES, Sun K, de Pee S, Kraemer K, Rah J-H, Moench-Pfanner R, et al. Relationship of maternal knowledge of anemia with maternal and child anemia and health-related behaviors targeted at anemia among families in Indonesia. Matern Child Health J. 2012 Dec;16(9):1913–25.  49.  WHO/CDC. Assessing the Iron Status of Populations. Second Edi. Geneva, Switzerland; 2004.  50.  Kassebaum NJ, Jasrasaria R, Naghavi M, Wulf SK, Johns N, Lozano R, et al. A systematic analysis of global anemia burden from 1990 to 2010. Blood. 2014 Jan 30;123(5):615–24.  51.  Clark S. Iron deficiency anemia. Nutr Clin Pract. 2008;23(2):128–41.  52.  Thurnham DI, Mccabe GP. Influence of infection and inflammation on biomarkers of nutritional status with an emphasis on vitamin A and iron. World Health Organization. Report: Priorities in the assessment of vitamin A and iron status in populations. Geneva, Switzerland; 2012.  53.  Douglas NM, Anstey NM, Buffet P a, Poespoprodjo JR, Yeo TW, White NJ, et al. The anaemia of Plasmodium vivax malaria. Malar J. BioMed Central Ltd; 2012 Jan;11:135.  54.  Masaninga F, Chanda E, Chanda-Kapata P, Hamainza B, Masendu HT, Kamuliwo M, et al. Review of the malaria epidemiology and trends in Zambia. Asian Pac J Trop Biomed. 2013 Feb;3(2):89–94.  95  55.  Kamuliwo M, Chanda E, Haque U, Mwanza-Ingwe M, Sikaala C, Katebe-Sakala C, et al. The changing burden of malaria and association with vector control interventions in Zambia using district-level surveillance data, 2006-2011. Malar J. 2013 Jan;12:437.  56.  Banda A. Prevalence of Schistosomiasis in Mbala District, Northern Province, Zambia. University of Zambia; 2006.  57.  Rutagwera DG. Malaria, Hookworms and Schistosomiasis Coninfection and Its Effect on Anemia Among School Children in Zambezi District of Zambia. University of Zambia; 2013.  58.  Barclay GPT, Charlesworth D, Lehmann H. Abnormal Haemoglobins in Zambia . A New Haemoglobin Zambia alpha60 (E9) Lysine to Asparagine. Britiish Med J. 1969;4:595–6.  59.  Muklwala EC, Banda J, Siziya S, Atenyi J, Fleming a F, Higgs DR. Alpha thalassaemia in Zambian newborn. Clin Lab Haematol. 1989 Jan;11(1):1–6.  60.  Van Dongen PWJ, van’t Hof MA. Sickle cell trait, malaria and anaemia in pregnant Zambian women. Trans R Soc Trop Med Hyg. 1983 Jan;77(3):402–4.  61.  Gayawan E, Arogundade ED, Adebayo SB. Possible determinants and spatial patterns of anaemia among young children in Nigeria: a Bayesian semi-parametric modelling. Int Health. 2014 Mar;6(1):35–45.  62.  Abubakar A, Uriyo J, Msuya SE, Swai M, Stray-Pedersen B. Prevalence and risk factors for poor nutritional status among children in the Kilimanjaro region of Tanzania. Int J Environ Res Public Health. 2012 Oct;9(10):3506–18.  63.  Wilunda C, Massawe S, Jackson C. Determinants of moderate-to-severe anaemia among women of reproductive age in Tanzania: analysis of data from the 2010 Tanzania demographic and health survey. Trop Med Int Health. 2013 Dec;18(12):1488–97.  64.  Meinzen-derr JK, Guerrero ML, Altaye M, Ortega-gallegos H, Ruiz-palacios GM, Morrow AL. Risk of Infant Anemia Is Associated with Exclusive Breast-Feeding and Maternal Anemia in a Mexican Cohort. J Nutr. 2006;136:452–8.  65.  Osei A, Pandey P, Spiro D, Nielson J, Shrestha R, Talukder Z, et al. Household food insecurity and nutritional status of children aged 6 to 23 months in Kailali District of Nepal. Food Nutr Bull. 2010;31(4):483–94.  66.  Quinn E a, Largado F, Power M, Kuzawa CW. Predictors of breast milk macronutrient composition in Filipino mothers. Am J Hum Biol. 2012;24(4):533–40.  67.  Kumar A, Rai AK, Basu S, Dash D, Singh JS. Cord blood and breast milk iron status in maternal anemia. Pediatrics. 2008 Mar;121(3):e673–7.  96  68.  Georgieff MK, Wewerka SW, Nelson C a, Deregnier R-A. Iron status at 9 months of infants with low iron stores at birth. J Pediatr. 2002 Sep;141(3):405–9.  69.  Pasricha S-R. Anemia: a comprehensive global estimate. Blood. 2014 Jan 30;123(5):611–2.  70.  Lozoff B, Georgieff M. Iron deficiency and brain development. Semin Pediatr Neurol. 2006;13:158–65.  71.  Lozoff B, Jimenez E, Wolf A. Long-term developmental outcome of infants with iron deficiency. N Engl J Med. 1991;  72.  Lozoff B, Jimenez E, Smith JB. Double Burden of Iron Deficiency in Infancy and Low Socioeconomic Status. Arch Pediatr Adolesc Med. 2006;160:1108–13.  73.  Stoltzfus RJ, Mullany L, Black RE. Iron deficiency anaemia. In: Ezzati M, Lopez AD, Rodgers A, Murray CJL, editors. Comparative Quantification of Health Risks: Global and Regional Burden of Disease Attribution to Selected Major Risk Factors. Geneva, Switzerland: World Health Organization; 2004. p. 163–209.  74.  Luo C, Mwela C, Campbell J. National baseline survey on prevalence and aetiology of anaemia in Zambia: a random cluster community survey involving children, women and men. Lusaka, Zambia; 1999.  75.  Eisele TP, Miller JM, Moonga HB, Hamainza B, Hutchinson P, Keating J. Malaria infection and anemia prevalence in Zambia’s Luangwa District: an area of near-universal insecticide-treated mosquito net coverage. Am J Trop Med Hyg. 2011 Jan;84(1):152–7.  76.  Mulendele E. Prevalence of Iron Deficiency Anemia in Children with Malaria. University of Zambia; 2007.  77.  Standing Committee on Nutrition. Zambia. Nutr Cris Situations. 2005;7.  78.  Van Dijk J, Hachaambwa L, Mulenga M, Mharakurwa S, Thuma P. Response of Haemoglobin Concentration to Oral Supplemental Iron in Children living ina Malaria endemic area of Zambia. Med J Zambia. 2007;34(2):86–91.  79.  Gibson RS, Kafwembe E, Mwanza S, Gosset L, Bailey KB, Mullen A, et al. A Micronutrient-Fortified Food Enhances Iron and Selenium Status of Zambian Infants but Has Limited Efficacy on Zinc. J Nutr. 2011;141:935–43.  80.  Lynch S. Case studies: iron. Am J Clin Nutr. 2011;94:673S–678S.  81.  Kung’u JK, Wright VJ, Haji HJ, Ramsan M, Goodman D, Tielsch JM, et al. Adjusting for the Acute Phase Response Is Essential to Interpret Iron Status Indicators among Young 97  Zanzibari Children Prone to Chronic Malaria and Helminth Infections. J Nutr. 2009;139:2124–31.  82.  Wander K, Brindle E, O’connor K a. Sensitivity and specificity of C-reactive protein and α(1) -acid glycoprotein for episodes of acute infection among children in Kilimanjaro, Tanzania. Am J Hum Biol. 2012;24(4):565–8.  83.  Thurnham D, McCabe L, Haldar S, Wieringa F, Northrop-Clewes C a, McCabe GP. Adjusting plasma ferritin concentrations to remove the effects of subclinical inflammation in the assessment of iron deficiency: a meta-analysis. Am J Clin Nutr. 2010;(4):546–55.  84.  Zimmermann MB, Molinari L, Staubli-Asobayire F, Hess SY, Chaouki N, Adou P, et al. Serum transferrin receptor and zinc protoporphyrin as indicators of iron status in African children. Am J Clin Nutr. 2005 Mar;81(3):615–23.  85.  Thomas C, Kobold U, Balan S, Roeddiger R, Thomas L. Serum hepcidin-25 may replace the ferritin index in the Thomas plot in assessing iron status in anemic patients. Int J Lab Hematol. 2011 Apr;33(2):187–93.  86.  Pfeiffer CM, Cook JD, Mei Z, Cogswell ME, Looker AC, Lacher D a. Evaluation of an automated soluble transferrin receptor (sTfR) assay on the Roche Hitachi analyzer and its comparison to two ELISA assays. Clin Chim Acta. 2007 Jul;382(1-2):112–6.  87.  Garrett DA, Sangha JK, Kothari MT, Boyle D. Field-friendly techniques for assessment of biomarkers of nutrition. 2011;94:685–90.  88.  Phiri KS, Calis JCJ, Siyasiya A, Bates I, Brabin B, Hensbroek MB Van. New cut-off values for ferritin and soluble transferrin receptor for the assessment of iron deficiency in children in a high infection pressure area. 2009;1103–6.  89.  Infusino I, Braga F, Dolci A, Panteghini M. Soluble transferrin receptor (sTfR) and sTfR/log ferritin index for the diagnosis of iron-deficiency anemia. A meta-analysis. Am J Clin Pathol. 2012 Nov;138(5):642–9.  90.  Verhoef H, West CE, Ndeto P, Burema J, Beguin Y, Kok FJ. Serum transferrin receptor concentration indicates increased erythropoiesis in Kenyan children with asymptomatic malaria. Am J Clin Nutr. 2001 Dec;74(6):767–75.  91.  Menendez C, Quinto LL, Kahigwa E, Alvarez L, Fernandez R, Gimenez N, et al. Effect of malaria on soluble transferrin receptor levels in Tanzanian infants. Am J Trop Med Hyg. 2001 Aug;65(2):138–42.  92.  Williams TN, Maitland K, Rees DC, Peto TE, Bowden DK, Weatherall DJ, et al. Reduced soluble transferrin receptor concentrations in acute malaria in Vanuatu. Am J Trop Med Hyg. 1999 May;60(5):875–8.  98  93.  Rees DC, Williams TN, Maitland K, Clegg JB, Weatherall DJ. Alpha thalassaemia is associated with increased soluble transferrin receptor levels. Br J Haematol. 1998 Nov;103(2):365–9.  94.  Uaprasert N, Rojnuckarin P, Bhokaisawan N, Settapiboon R, Wacharaprechanont T, Amornsiriwat S, et al. Elevated serum transferrin receptor levels in common types of thalassemia heterozygotes in Southeast Asia: a correlation with genotypes and red cell indices. Clin Chim acta. Elsevier B.V.; 2009 May;403:110–3.  95.  Singhal A, Cook JD, Skikne BS, Thomas P, Serjeant B, Serjeant G. The clinical significance of serum transferrin receptor levels in sickle cell disease. Br J Haematol. 1993 Jun;84(2):301–4.  96.  Punnonen K, Irjala K, Rajamäki a. Serum transferrin receptor and its ratio to serum ferritin in the diagnosis of iron deficiency. Blood. 1997 Feb 1;89(3):1052–7.  97.  Cook J, Flowers C, Skikne B. The quantitative assessment of body iron. Blood. 2003;3359–63.  98.  Baeten JM, Richardson B a, Bankson DD, Wener MH, Kreiss JK, Lavreys L, et al. Use of serum retinol-binding protein for prediction of vitamin A deficiency: effects of HIV-1 infection, protein malnutrition, and the acute phase response. Am J Clin Nutr. 2004 Feb;79(2):218–25.  99.  WHO. Serum ferritin concentrations for the assessment of iron status and iron deficiency in populations. Geneva, Switzerland; 2011.  100.  Hayes AF, Cai L. Using heteroskedasticity-consistent standard error estimators in OLS regression: an introduction and software implementation. Behav Res Methods. 2007 Nov;39(4):709–22.  101.  Tabachnick BG, Fidell LS. Using Multivariate Statistics. 5th Editio. Needham Heights, MA, USA: Allyn & Bacon Incorporated; 2006.  102.  Baker RD, Greer FR. Diagnosis and prevention of iron deficiency and iron-deficiency anemia in infants and young children (0-3 years of age). Pediatrics. 2010 Nov;126(5):1040–50.  103.  Northrop-Clewes C a, Thurnham DI. Biomarkers for the differentiation of anemia and their clinical usefulness. J Blood Med. 2013 Jan;4:11–22.  104.  Chanda E, Kamuliwo M, Steketee RW, Macdonald MB, Babaniyi O, Mukonka VM. An Overview of the Malaria Control Programme in Zambia. ISRN Prev Med. 2013;2013:1–8.  99  105.  Wamani H, Astrøm AN, Peterson S, Tumwine JK, Tylleskär T. Boys are more stunted than girls in sub-Saharan Africa: a meta-analysis of 16 demographic and health surveys. BMC Pediatr. 2007 Jan;7:17.  106.  Siekmans K, Receveur O, Haddad S. Can an integrated approach reduce child vulnerability to anaemia? Evidence from three african countries. PLoS One. 2014 Jan;9(3):e90108.  107.  Siegel EH, Stoltzfus RJ, Khatry SK, Leclerq SC, Katz J, Tielsch JM. Epidemiology of anemia among 4- to 17-month-old children living in south central Nepal. Eur J Clin Nutr. 2006 Feb;60(2):228–35.  108.  Agho KE, Dibley MJ, D’Este C, Gibberd R. Factors associated with haemoglobin concentration among Timor-Leste children aged 6-59 months. J Health Popul Nutr. 2008 Jun;26(2):200–9.  109.  Osório MM, Lira PIC, Ashworth A. Factors associated with Hb concentration in children aged 6-59 months in the State of Pernambuco, Brazil. Br J Nutr. 2004 Feb;91(2):307–15.  110.  Checkley W, Gilman RH, Black RE, Epstein LD, Cabrera L, Sterling CR, et al. Effect of water and sanitation on childhood health in a poor Peruvian peri-urban community. Lancet. 2004 Jan 10;363(9403):112–8.  111.  Desai MR, Terlouw DJ, Kwena AM, Phillips-Howard P a, Kariuki SK, Wannemuehler K a, et al. Factors associated with hemoglobin concentrations in pre-school children in Western Kenya: cross-sectional studies. Am J Trop Med Hyg. 2005 Jan;72(1):47–59.  112.  Oski FA. Iron deficiency in infancy and childhood. N Engl J Med. 1993;329(3):190–3.  113.  Mburu ASW, Thurnham DI, Mwaniki DL, Muniu EM, Alumasa F, de Wagt A. The influence and benefits of controlling for inflammation on plasma ferritin and hemoglobin responses following a multi-micronutrient supplement in apparently healthy, HIV+ Kenyan adults. J Nutr. 2008 Mar;138(3):613–9.  114.  Hipgrave DB, Fu X, Zhou H, Jin Y, Wang X, Chang S, et al. Poor complementary feeding practices and high anaemia prevalence among infants and young children in rural central and western China. Eur J Clin Nutr. 2014 Aug;68(8):916–24.  115.  Faber M. Dietary intake and anthropometric status differ for anaemic and non-anaemic rural South African infants aged 6-12 months. J Health Popul Nutr. 2007 Sep;25(3):285–93.  116.  Knowles J, Thurnham DI, Phengdy B, Houamboun K, Philavong K, Keomoungkhone I, et al. Impact of inflammation on the biomarkers of iron status in a cross-sectional survey of Lao women and children. Br J Nutr. 2013 Dec;110(12):2285–97.  100  117.  Siekmann J, Allen L, Bwibo N, Demment M, Murphy S, Neumann C. Kenyan school children have multiple micronutrient deficiencies, but increased plasma vitamin B-12 is the only detectable micronutrient response to meat or milk supplementation. J Nutr. 2003;133:3972S–3980S.    101  Appendix A: List of Selected Zones, by Catchment Area Kaka Catchment Area: Chisafwa, Chisenga, Jeremia Lutundu, Maloni, Mwabala Kamuzwazi Catchment Area: Chisitu, Kachenyente, Kakone, Kawala, Mutipe Kawimbe Catchment Area: Chiwindi, Chupa, Kawama, Mfwambo, Mulalo, Mulefu Mambwe Mission Catchment Area: Chilinde, Ilondola, Mambwe Mission Static, Mutonga, Nsindano Mbala Urban Catchment Area: Chila View, Chulungoma, Lunzua, Mbala Urban Static, Ndundundu, Zombe Mpande Catchment Area: Chikunta, Chindo, Kapufi, Mpande Static, Nzelani, Satoka Nondo Catchment Area: Chikoti, Chomba, Kanyambi, Mwiluzi, Nondo Static, Shikulu Musonda Senga Catchment Area: Kalinda, Kapoli, Lusale, Mipangala, Musombizi, Muzizi Tulemane: Isanya, Kamyanga, Londe, Maround, Musipazi, St Paul  102  Appendix B: Consent Forms Control Group Name of Investigators: Mrs Agnes Aongola and Ms Mélanie Suter Study Sponsors: UNICEF and Irish Aid Participants: Mothers of children 6-11 months (at time of entry) Part I: Information Sheet Purpose Many children in Zambia are not receiving adequate nutrients from food for proper growth and development. We are most interested in a nutrient called iron and its inadequate intake in children 6-23 months which can cause a condition called anaemia. Anaemia is very common among young children in Zambia, partially because the foods they eat contain very little iron. We will use the information gathered during this study to inform the scaling up of the distribution of nutritional supplements. Procedure If you agree that you and your child should participate, you will be provided with a mosquito net and deworming tablets (mebendazole) for your child. You will be asked to answer some questions at the beginning of the study and a blood sample will be collected from your child to measure how much of various micronutrients are in the child’s blood and to measure the amount of blood (Haemoglobin). We will also take height and weight measurements of your child. Then, a follow up visit will be done after 6 months and 12 months to collect additional information on your child’s health, and collect another blood sample to see how your child’s health has improved. Benefits Preventing anaemia in children aged 6-23 months of age can have long term benefits on a child’s growth and development. The mosquito net will help protect your child from malaria and other diseases and the de-worming tablet will cure you child of worms. Further, if your child is found to be severely anaemic or severely malnourished, we will provide a referral for treatment at the nearest health clinic. Risks and discomfort  There are no major risks associated with participation in this study. The blood drawing is momentarily uncomfortable. All blood draws will be performed by a trained phlebotomist or qualified health care professional with extensive experience working with children. Confidentiality The information that you provide us will be kept strictly confidential. Your privacy and your information are respected. Personal information will not be shared with anyone other than the research team. Costs/Compensation You will not be charged for your participation in the research. All costs associated with the study will be paid by the sponsors. You and your child can benefit from this program, but will not receive any payment to take part in the research. Complaints Should you wish to raise any ethical issues, you may be concerned about please contact the chairman of the TDRC Ethics committee Dr T G Ngulube at Box 320168, Lusaka, phone number 0955914844. Right to refuse or withdraw  You may choose not to participate in this study or have your child participate in the study. You and your child may stop participating in the study at any time that you wish.       103   Part II: Certificate of Consent I have been asked to give consent for my daughter/son to participate in the research study entitled “MNP pilot study for Development of a Home Fortification Program for Young Children in Zambia” which will involve providing information related to my child’s health and to the intervention and collecting blood samples on my child three times during the study. I have read/or it has been read to me and understood the forgoing information. I have had the opportunity to ask questions and I was given satisfactory answers to my questions. Name of child: ______________________________ Name of mother/father: ______________________________ Signature of mother/father: ______________________________ Signature of presenter: ______________________________ Date: ______________________________ If illiterate: A literate witness must sign and parents who are illiterate should include their thumb print as well. I have witnessed the accurate reading of the consent from the parent of the potential participant and the individual has had the opportunity to ask questions. I confirm that the individual has given consent freely. Signature of witness: ________________________________  Thumbprint  104  Intervention Group Name of Investigators: Mrs Agnes Aongola and Ms Mélanie Suter Study Sponsors: UNICEF and Irish Aid Participants: Mothers of children 6-11 months (at time of entry) Part I: Information Sheet Purpose Many children in Zambia are not receiving adequate nutrients from food for proper growth and development. We are most interested in a nutrient called iron and its inadequate intake in children 6-23 months which can cause a condition called anaemia. Anaemia is very common among young children in Zambia, partially because the foods they eat contain very little iron. We will use the information gathered during this study to inform the scaling up of the distribution of nutritional supplements. Procedure If you agree that you and your child should participate, your child will be provided with a nutritional supplement for 30 days every two months for a 1-year period. The nutritional supplement is a powder form of iron and other important vitamins and minerals, contained in small packets. You will be expected to add the nutritional supplement to your child’s food and feed it to them over the duration of the pilot study. You will also be provided with a mosquito net and de-worming tablets (mebendazole) for your child. After the start of the program you will be given oral and written instructions on how to prepare and serve the nutritional supplement. You will be asked to answer some questions at the beginning of the study and a blood sample will be collected from your child to measure how much of these micronutrients are in the child’s blood and to measure the amount of blood (Haemoglobin). We will also take height and weight measurements of your child. Then, a follow up visit will be done after 6 months and 12 months to collect additional information on your child’s health and your experience with the nutritional supplement and collect another blood sample to see how your child’s health has improved. Benefits Preventing anaemia in children aged 6-23 months of age can have long term benefits on a child’s growth and development. The mosquito net will help protect your child from malaria and other diseases and the de-worming tablet will cure you child of worms. Further, if your child is found to be severely anaemic or severely malnourished, we will provide a referral for treatment at the nearest health clinic. Risks and discomfort There are no major risks associated with participation in this study. The blood drawing is momentarily uncomfortable. All blood draws will be performed by a trained phlebotomist or qualified health care professional with extensive experience working with children. Confidentiality The information that you provide us will be kept strictly confidential. Your privacy and your information are respected. Personal information will not be shared with anyone other than the research team. Costs/Compensation You will not be charged for your participation in the research. All costs associated with the study will be paid by the sponsors. You and your child can benefit from this program, but will not receive any payment to take part in the research.     105  Complaints Should you wish to raise any ethical issues, you may be concerned about please contact the chairman of the TDRC Ethics committee Dr T G Ngulube at Box 320168, Lusaka, phone number 0955914844. Right to refuse or withdraw You may choose not to participate in this study or have your child participate in the study. You and  your child may stop participating in the study at any time that you wish.   Part II: Certificate of Consent I have been asked to give consent for my daughter/son to participate in the research study entitled “MNP  pilot study for Development of a Home Fortification Program for Young Children in Zambia” which will  involve my child taking a nutritional supplement for 30 days every two months for a year, providing  information related to my child’s health and to the intervention and collecting blood samples on my child  three times during the study. I have read/or it has been read to me and understood the forgoing information. I have had the opportunity to ask questions and I was given satisfactory answers to my questions. Name of child: ______________________________ Name of mother/father: ______________________________ Signature of mother/father: ______________________________ Signature of presenter: ______________________________ Date: ______________________________ If illiterate: A literate witness must sign and parents who are illiterate should include their thumb print as well. I have witnessed the accurate reading of the consent from the parent of the potential participant and the individual has had the opportunity to ask questions. I confirm that the individual has given consent freely. Signature of witness: ________________________________  Thumbprint 106  Appendix C: Questionnaire  Module 0: Identification Information Q001. Child ID: Q006. Catchment Area: Q002. Child’s Name: Q007. Zone: Q003. Child’s Birthday*: _ _ /_ _ /_ _ _ _ (dd/mm/yyyy) Q008. Village: Q004. Mother’s Name: Q009. Interviewer Code: Q005. Date of Interview: Q010. Field Editor Code: * CHECK UNDER 5 CARD TO VERIFY. IF CHILD IS YOUNGER THAN 6 MONTHS OR OLDER THAN 12 MONTHS IT IS NOT ELIGIBLE FOR THIS STUDY. IF EXACT DATE OF BIRTH IS NOT KNOWN ONLY RECORD MONTH.  INTERVIEWER: Explain to interviewee what you are going to measure, how and why.  Module 1: Blood Sampling, Anthropometry & Oedema Blood sample  Q101. Blood sample taken? IF ANSWER IS NO, RECORD REASON. 0= No  1= Yes Reason _____________________________  Q102. Haemoglobin count _ _._  MUAC  Q103. First measurement MUAC _ _._ cm  107  Q104. Second measurement MUAC _ _._ cm  Q105. Third measurement (only record, if difference between first & second measurement is more than 5mm) MUAC _ _._ cm  Q106. Does the child have a MUAC below 11.5 cm? 0= No  1= Yes  Weight (kg)   Q107.  Weight of mother and child together  ONLY WEIGH ONCE  _ _ _ . _  kilograms   Q108. Weight of mother alone  ONLY WEIGH ONCE _ _ _ . _  kilograms  Q109. Weight of child alone  CALCULATE WITH CALCULATOR AT ANTHRO STATION _ _ _ . _  kilograms  Length (cm) Q110. First measurement  _ _. _   cm  Q111. Second measurement _ _ . _  cm  Q112. Third measurement (only record, if difference between first & second measurement is more than 5mm) _ _ . _  cm  108  Q113. Does the child have a weight-for-length z-score below -3 SD?  0= No  1= Yes  Oedema Check Q114. Bilateral oedema present? 0= No  1= Yes  SAM Q115. Child diagnosed with SAM based on Anthropometry (Check: Q106, Q113 or Q114)? 0= No  1= Yes  End interview. Immediately inform field editor.  INTERVIEWER: Thank you for meeting with us today and participating in this study. There are no right or wrong answers to the questions. We are here to learn from you so we can best meet the needs of families in the future. If you cannot or do not want to answer any of the questions just let us know and we will go on to the next question. We really appreciate your time and your participation. May we start the interview now? Module 2: Household Information Q201. How old are you? Age of primary caregiver in years  Q202. How many people in total currently live in your household? DEFINITION: ALL PEOPLE WHO HAVE LIVED IN THE HOUSE FOR THE PAST 6 MONTHS AND INTEND TO STAY IN THE HOUSEHOLD. Number of people living in the household  109  Q203. How many children are in your primary care? Number of children  Q204. Please tell me the sex and ages of all your children under five starting with your youngest child.  FOR EACH CHILD, STARTING WITH THE YOUNGEST, CIRCLE THE APPLICABLE NUMBER (1 OR 2) INDICATING SEX AND RECORD AGE (IN MONTHS) IN BOXES. 1. Child 1= Male  /  2= Female  Age in months 2. Child 1= Male  /  2= Female  Age in months 3. Child 1= Male  /  2= Female  Age in months 4. Child 1= Male  /  2= Female  Age in months 5. Child 1= Male  /  2= Female  Age in months 6. Child 1= Male  /  2= Female  Age in months 7. Child 1= Male  /  2= Female  Age in months        Q205.  What is your relationship to the child who is enrolled in this study? 1= Birth mother 99= Other – Specify: ________________________________  Q206. Have you ever attended school? 0= No  skip to Q208. 1= Yes   Q207. If yes, what is the highest level of school you attended? 1 = Primary school (Grade 1-7) 2 = Secondary school (Grade 8-12)  3 = Tertiary/higher education (Certificate-PhD) 99 = Other (exclude formal school) – Specify: ________________________________________  110  Q208. Do you have any income generating activities in your household (e.g. cash crops, crafts, etc.)?  0= No  Skip to Q210. 1= Yes 2= Sometimes  Q209. If yes, what is your main income generating activity in the household?        1= Farming 2= Fishing 3= Agricultural piece work 4= Non-Agricultural piece work 5 = Service/Salaried worker 6 = Business/Traders 99 = Other – Specify: _______________________________   Q210. Does your household have any agricultural land for cultivating (e.g. maize, cassava, cabbage, beans, and groundnuts)?  0= No  Skip to Q212. 1 = Yes   Q211. Which crop(s) did you grow on this plot in the last season? RECORD ALL GIVEN ANSWERS BY RECORDING THE LETTER IN THE CORRESPODNING BOX. DO NOT READ. PROMPT FOR MORE. a. Cereals (e.g. maize, wheat, rice, millet) b. Legumes (beans, lentils, etc.)  c. Vegetables (green leafy and others) d. Roots (potatoes, cassava, sweet potatoes, etc.) e. Fruit – Specify: _____________________________ f. None g. Other–Specify: _____________________________       111  Q212. Does your family have a fruit or vegetable garden? 0= No  Skip to Q215. 1= Yes  Q213. What do you grow in your garden vegetables, fruit, or both?  1= Vegetables only 2= Fruits only 3= Vegetables & fruits  Q214. What is the main use of fruits or vegetables grown in your garden? 1= Home consumption 2= Sell 3= Give to others 99= Other – Specify: ________________________________   Q215. Do you own any farm animals? 0= No   Skip to Q218. 1= Yes  Q216. If yes, what type of farm animals do you own? RECORD ALL GIVEN ANSWERS BY RECORDING THE LETTER IN THE CORRESPODNING BOX. DO NOT READ. PROMPT FOR MORE. h. Cow i. Goat j. Chicken k. Pig l. Rabbit  m. Sheep n. Donkey h. Other—Specify: ______________________________        112  Q217. What is the main use of the farm animal(s) that you own? RECORD ALL GIVEN ANSWERS BY RECORDING THE LETTER IN THE CORRESPODNING BOX. DO NOT READ. PROMPT FOR MORE.  a. Home consumption b. Sell byproducts (eg. Manure, milk, eggs) c. Sell entire animal d. Give the by-products to others (e.g. Manure, milk, eggs) e. Give entire animal to others f. Use as work animal(s)/transport g. Other—Specify: _____________________________        Q218. Are there months when you or your family have to skip meals (i.e. don’t have enough food for everyone)? 0= No  Skip to Module 3 1= Yes  Q219. If yes, during which months?  ___________________________________________    Module 3: Drinking Water, Hygiene & Sanitation Q301. What is the main source of drinking water for your household in the dry season? PLEASE RECORD CODE IN BOXES. 1= Piped water (into dwelling/compound/plot) 2= Public tap (e.g.borehole) 3= Open/unprotected/shallow/traditional public well 4= Protected (including cover) public well  5= Pond/river/streams /lake/spring 6= Water tank 7= Bottled water 99= Other -Specify:___________________________  Q302. What is the main source of drinking water for your household in the rainy season? PLEASE RECORD CODE IN 1= Piped water (into dwelling/compound/plot) 2= Public tap (e.g. borehole)  3= Open/unprotected/shallow/traditional public well 4= Protected (including cover) public well   113  BOXES.   5= Pond/river/streams/lake/spring 6= Rainwater harvesting 7= Water tank 8= Bottled water 99= Other – Specify: _______________________________   Q303. Do you treat water in any way to make it safer to drink? 0 = No  Skip to Q305. 1 = Yes  Q304. If yes, what do you usually (most often) do to the water to make it safer to drink? PLEASE RECORD CODE IN BOXES. ONLY RECORD ONE OPTION  WHAT IS MOST OFTEN DONE.   1= Boil  2= Add bleach/chlorine 3= Strain it through a cloth 4= Use water filter (ceramic, sand, composite, etc.) 5= Solar disinfection 6= Let it stand and settle  99= Other–Specify: __________________________  Q305. Do you wash your hands with soap or ash? 0= No  Skip to Q308. 1= Yes  Q306. If yes, which do you most often use, soap or ash? 1= Soap 2= Ash  Q307. If yes, when do you wash your hands with soap/ash? RECORD ALL GIVEN ANSWERS BY RECORDING a. Before preparing/handling food  b. After preparing food c. Before feeding children d. Before eating e. After eating f. After field work/cleaning   114  THE LETTER IN THE CORRESPODNING BOX. PROMPT FOR MORE. DO NOT READ ANSWERS ALOUD.  g. After changing babies/cleaning child who has defecated  h. After defecating/using toilet facility i. While washing rest of body j. Other – Specify: ______________________________         Q308. Do you wash your children’s hands with soap or ash? 0 = No     Skip to Q311. 1 = Yes  Q309. If yes, which do you most often use, soap or ash? 1= Soap 2= Ash    115  Q310. If yes, when do you wash your children’s hands with soap or ash? RECORD ALL GIVEN ANSWERS BY RECORDING THE LETTER IN THE CORRESPODNING BOX. PROMPT FOR MORE. DO NOT READ ANSWERS ALOUD. a. Before they eat b. After they eat c. After they defecate/use toilet facility d. While washing rest of the body e. Other – Specify:_________________________      Q311. What kind of toilet facility do members of your household usually (most often) use? PLEASE RECORD CODE IN BOXES.  1= Flush/pour flush toilet 2= Ventilated improved pit latrine (VIP) 3= Pit latrine with slab 4= Pit latrine without slab (traditional latrine) 5= Bucket 6= No facilities/bush/field/river    Skip to Q314. 99=Other–Specify: ___________________________  Q312. Do you share this facility with other households? 0 = No     Skip to Q314. 1 = Yes  Q313. How many other households share this toilet facility? Number of households 77= Don’t know     116  Q314. The last time (NAME) passed stools, what was done to dispose of the stool? PLEASE RECORD CODE IN BOXES.  1= Child used toilet/latrine 2= Put/rinsed into toilet or latrine 3= Put/rinsed into drain or ditch 4= Thrown into garbage 5= Buried  6= Left in the open/river/lake 99=Other–Specify: ___________________________   Module 4: Early Childhood Development Q401. Do you know how a child could be stimulated to help him/her learn new things? DO NOT PROVIDE EXAMPLES 0= No    Skip to Q403. 1= Yes 77= Don't Know  Skip to Q403.  Q402. If yes, can you tell me how a child can be stimulated? PROBE TO LEARN SPECIFICALLY HOW THE INTERVIEWEE WOULD STIMULATE A CHILD.   1. _________________________________________  _________________________________________  2. _________________________________________  _________________________________________  3. _________________________________________  _________________________________________  Q403. How many children’s books or picture books do you have for (NAME)? Number of books  Q404a. I am interested in learning about the things that (NAME) plays with when he/she is at home. Does (NAME) play with: READ OUT OPTIONS. RECORD ALL GIVEN ANSWERS BY RECORDING THE LETTER IN THE CORRESPODNING BOX. a. Homemade toys (e.g. dolls, cars, or other toys made at home) b. Toys from a shop or manufactured toys c. Household objects (e.g. bowls, pots) or objects found outside (e.g. sticks, rocks, animal shells, leaves)  d. No toys/child is too young to play with toys  Skip to Q405a.     117  Q404b. DEPENDING ON ANSWER GIVEN ABOVE, PROBE TO LEARN SPECIFICALLY WHAT THE CHILD PLAYS WITH TO ASCERTAIN THE RESPONSE.  Homemade toys ___________________________________  _________________________________________________  Toys from a shop __________________________________  _________________________________________________  Household or outside objects ________________________  _________________________________________________  Q405a. Sometimes adults taking care of children have to leave the house to work in the field, go shopping, wash clothes, or for other reasons and have to leave young children. On how many days in the past week was (NAME) left alone for more than an hour? Number of days left alone for more than one hour 77= Don't Know  Q405b. On how many days in the past week was (NAME) left in the care of someone less than 10 years old, for more than an hour? Number of days left with other child for more than an hour 77= Don't Know    118  Q406. In the past 3 days, did you or any household member over 15 years of age engage in any of the following activities?  READ OUT OPTIONS. CIRCLE LETTERS ACCORDING TO ANSWER: a. Read books to or looked at pictures with (NAME) b. Told stories to (NAME) c. Sang songs to (NAME) or with (NAME), including lullabies d. Took (NAME) outside the home, compound, yard or enclosure e. Played with (NAME) f. Named, counted, or drew things to or with (NAME)                                        Mother     Father     Other    No one a. Read books    A  B  C  D b. Told stories    A  B  C  D c. Sang songs    A  B  C  D d. Took outside    A  B  C  D e. Played with    A  B  C  D f. Named/counted    A  B  C  D   INTERVIEWER: I would now like to ask you some questions about the health of (NAME). Module 5: Child Health & Health Seeking Behaviour Q501. Has (NAME) had diarrhea, (this means 3 or more times a day of loose stools) in the past 2 weeks? 0= No 1= Yes 77= Don't Know  Q502. Has (NAME) been ill with a cough at any time in the past 2 weeks? 0= No 1= Yes 77= Don't Know  Q503. Has (NAME) been ill with a fever at any time in the past 2 weeks? 0= No  Skip to Q507. 1= Yes 77= Don't Know  119  Q504. Is (NAME) still sick with a fever?   0= No  1= Yes 77= Don't Know  Q505. Did you seek advice or treatment for the fever from any source?  0= No  Skip to Q507. 1= Yes  Q506. Did (NAME) receive a finger prick or heal prick to test the fever/illness?  0= No  1= Yes  Q507. Is (NAME) currently treated for malaria?  0= No 1= Yes  Q508. During illness of (NAME), do you generally seek advice or treatment?  0= No  Skip to Q510. 1= Yes  Q509. If yes, from whom do you generally seek advice/treatment?  RECORD ALL GIVEN ANSWERS BY RECORDING THE LETTER IN THE CORRESPODNING BOX. DO NOT READ ALOUD. PROMPT FOR MORE. Health Sector a. Hospital b. Health centre/health post c. Outreach point/mobile clinic d. Drug sellers/Pharmacy  e. Other – Specify: _________________________ Community f. Midwife g. CHW/TBA/volunteer h. Traditional healer i. Mother support group j. Friend or neighbor k. Mother or mother-in-law l. Other relative  m. Other – Specify: _________________________               Q510. Has (NAME) received a Vitamin A capsule within the past 6 months? 0= No 1= Yes 77= Don't Know  120  Q511. Has (NAME) received a de-worming tablet within the past 6 months? 0= No 1= Yes 77= Don't Know  Q512. Have you been tested for HIV during your pregnancy with (NAME)?  0= No 1= Yes  Q513. Has (NAME) been tested for HIV?  0= No  skip to Q515. 1= Yes  Q514. What is (NAME)’s HIV status? 0= Negative 1= Positive 77= Don't Know  Q515. Does (NAME) currently suffer from tuberculosis?  0= No 1= Yes 77= Don't Know  Q516. Does (NAME) sleep under a mosquito net? 0= No 1= Yes   Module 6: Mother’s Knowledge on IYCF & Anaemia Q601. Have healthcare providers ever talked to you about breastfeeding or how to feed (NAME)? 0= No  skip to Q607. 1= Yes   Q602. In the last two months, how many counselling sessions at an outreach point or health centre have you attended? Record number of times  121  Q603. Can you tell me what the topics of those counselling sessions were?  PROBE AND RECORD AS MUCH DETAIL AS YOU CAN. Topic ___________________________________________  _________________________________________________  Topic ___________________________________________  _________________________________________________  Q604. Who was conducting the counselling sessions you attended?   1= Healthcare provider 2= Community volunteer 3= Others – Specify? _______________________________  Q605. Were you provided with brochures or materials to take home with you? 0= No  skip to Q607. 1= Yes   Q606. If yes, what types of brochures did you receive?  RECORD ALL GIVEN ANSWERS BY RECORDING THE LETTER IN THE CORRESPODNING BOX. DO NOT READ ALOUD. PROMPT FOR MORE. a. Breastfeeding brochure b. Complementary feeding brochure c. Maternal nutrition brochure d. Other – Specify? ____________________________      122  Q607. Can you tell me how long after birth a child should start receiving breast milk/be put to the breast? PLEASE RECORD CODE IN BOXES TO FAR RIGHT. DEPENDING ON ANSWER ALSO RECORD NUMBER OF HOURS OR DAYS. 1= Immediately (within one hour) 2= Hours  Record number of hours Q607a. 3= Days   Record number of days   Q607b. 77= Don’t know  Q608. Can you tell me until what age a baby should receive only breast milk, that is, no other food, water or other fluids? Record months 77= Don’t know  Q609. At what age should a child receive something else than breast milk such as enriched porridge? Record months 77= Don’t know  Q610. Have you ever heard of a condition called anaemia or iron deficiency (shortage of blood)? IF MOTHER DOES NOT KNOW, EXPLAIN WHAT IT IS WITHOUT NAMING SYMPTOMS FROM Q614. IF MOTHER STILL DOES NOT KNOW SKIP TO Q615. OTHERWISE PROCEED WITH Q611.  0= No  FOLLOW INSTRUCTION. 1= Yes   123  Q611. Do you know what the causes of anaemia or iron deficiency are? 0= No  skip to Q611. 1= Yes   Q612. If yes, what are the causes of anaemia or iron deficiency? RECORD ALL GIVEN ANSWERS BY RECORDING THE LETTER IN THE CORRESPONDING BOX. PROMPT FOR MORE. DO NOT READ ANSWERS ALOUD.  a. Malaria b. Hookworms   c. Infection d. Bleeding e. Dietary (lack of iron in diet) f. Other – Specify: _______________________________      Q613. Can you identify some signs that a child has anaemia?  0= No  skip to Q615. 1= Yes   Q614. If yes, please tell me some of the signs that a child has anaemia or a shortage of blood?  RECORD ALL GIVEN ANSWERS BY RECORDING THE LETTER IN THE CORRESPONDING BOX. DO NOT READ ANSWERS ALOUD. PROMPT FOR MORE.   a. White/pale skin, eyes, nails, hands b. Feeling faint, dizziness  c. Apathy/fatigue (child doesn’t play) d. Oedema e. Weight loss f. Fever/high temperature  g. Other – Specify: _______________________________       Q615. Do you know of any foods that are high in iron? 0= No  skip to Module 7. 1= Yes   124  Q616. If yes, please tell me what foods are high in iron?  RECORD ALL GIVEN ANSWERS BY RECORDING THE LETTER IN THE CORRESPONDING BOX. PROMPT FOR MORE. DO NOT READ ANSWERS ALOUD.  a. Organ meats (e.g. liver, kidney, heart) b. Meat (e.g. beef, pork) or poultry (e.g. chicken) c. Fresh or dried fish (e.g. kapenta) d. Green leafy vegetables (e.g. rape, chibwabwa, katapa) e. Other vegetables or fruit (e.g. oranges, tomatoes) f. Legumes (e.g. beans, lentils) g. Tubers (e.g. potatoes) h. Eggs, milk i. Others – Specify: ______________________________         INTERVIEWER: The next module is for the child in the household between 6-11 months who will be participating in the pilot study. I would like to ask you some questions about the feeding of (NAME) including what liquids and foods he/she consumed yesterday. Module 7: Infant & Young Child Feeding Practices & Food Consumption Part I: Breastfeeding & Feeding Practices  Q701. Did you ever breastfeed (NAME)? 0= No  Skip to Q703. 1= Yes  Q702. How long after delivery, did you start breastfeeding (NAME)? PLEASE RECORD CODE IN BOXES TO FAR RIGHT. DEPENDING ON ANSWER ALSO RECORD NUMBER OF HOURS OR DAYS. 1= Immediately (within the first hour) 2= Hours  Record number of hours Q702a.  3= Days    Record number of days   Q702b.  77= Don’t know  125  Q703. In the first three days after delivery was (NAME) given anything to drink other than breast milk, such as other milk, tea, traditional medicine, sugar water or plain water? 0= No 1= Yes 77= Don't Know  Q704. Are you still breastfeeding (NAME)? 0= No 1= Yes  Skip to Q706.  Q705. If no, for how many months did you breastfeed (NAME)? Months 77= Don’t know   Q706. Did (NAME) drink anything from a bottle with a nipple yesterday or last night? 0= No 1= Yes  Q707. Has (NAME) started receiving any semi-solid or mashed foods prepared specifically for the baby? 0= No  Skip to Q709. 1= Yes   Q708. If yes, at what age did (NAME) start receiving semi-solid or mashed foods prepared specifically for the baby? Months 77= Don’t Know  Q709. Has (NAME) received any foods from the family pot without being mashed or prepared specifically for the baby? 0= No  Skip to Q717. 1= Yes  Q710. If yes, at what age did (NAME) start receiving foods from the family pot? Months  77= Don’t know  Q711. Does (NAME) eat from his/her own separate bowl?   0= No  Skip to Q713. 1= Yes   Q712. a) If yes, at what age did (NAME) start eating from his/her own separate bowl?  b) At what age will (NAME) stop eating from his/her own separate bowl? a) Record age in months b) Record age in months bb)  Others–Specify: __________________________   126  Q713. Does (NAME) usually (most often) feed him/herself or does someone help put the food in (NAME’s) mouth? 1= (NAME) feeds herself/himself 2= Someone else feeds him/her)  Skip to Q715.  Q714. At what age did (NAME) start feeding him or herself?   Skip to Q716. after recording answer  Months 77= Don’t Know  Q715. Who usually (most often) feeds (NAME)?  PLEASE RECORD CODE IN BOXES. IF CAREGIVER SAYS OLDER SIBLING FEEDS CHILD, RECORD AGE OF SIBLING. 1= Myself 2= Father/Husband  3= Grandparent 4= Older sibling (RECORD AGE IN YEARS) Q715b.  5= Neighbour  99= Other – Specify: _______________________________  Q716. Do you/they encourage (NAME) to eat? 0= No 1= Yes    Part II: Food Consumption   Q717. Now I would like to ask you about some liquids that (NAME) may have had at any time yesterday during the day or night. Did (NAME) have any of the following? READ OUT THE WHOLE LIST. RECORD ALL GIVEN ANSWERS BY RECORDING THE LETTER IN THE CORRESPODNING BOX.   a. Breast milk?  b. Plain water? c. Infant formula? d. Milk such as tinned, powdered, or fresh animal milk? e. Sweetened water, juice, fruit juice, soda (carbonated drinks)? f. Soup broth?  g. Dilute maize porridge/beverage h. Any other liquids (e.g. tea, coffee)? – Specify:  _______________________________________          127  Q718. How many times at any time yesterday during the day or night did you give (NAME) infant formula or non-human milk (tinned, powdered, or fresh animal milk)? IF Q707. AND Q709. WERE ANSWERED WITH ‘NO’  SKIP TO THE END OF THE INTERVIEW.  Number of times     Q719.  Now I am going to ask you about the foods (NAME) consumed yesterday whether at home or outside the home.  Please describe everything that (NAME) ate yesterday during the day or night.  PROBE AND GET AS MUCH DETAIL AS POSSIBLE. IF RESPONDENT MENTIONS MIXED DISHES LIKE PORRIDGE, SAUCE OR RELISH, PROBE: WHAT INGREDIENTS WERE IN THAT DISH? RECORD CODE IN BOXES ACCORDINGLY.   Q719a. Porridge e.g. maize, cassava, mixed grains 0= No  skip to Q719c. 1= Yes  Q719b. If yes, what was the thickness? PLEASE RECORD CODE IN BOX. 1= Thin – like liquid soup/munkoyo 2= Medium – not thick enough to stay on a spoon 3= Thick – thick enough to stay on a spoon  Q719c. Maize, bread, wheat, rice, sorghum, millet or other staple foods made from grain 0= No 1= Yes  Q719d. White potatoes, white sweet potatoes, cassava or other white root vegetables 0= No 1= Yes  128  Q719e. Beans, lentils, cowpeas, soya, groundnuts  0= No 1= Yes  Q719f. Milk, yogurt, cheese or other foods made from milk 0= No 1= Yes  Q719g. Liver, kidney, heart, other organ meats 0= No 1= Yes  Q719h. Meat (beef, pork, lamb, goat), poultry (chicken or other birds), caterpillar PROBE WHETHER CHILD HAS EATEN ACTUAL MEAT OR ONLY GRAVY. 0= No 1= Yes  Q719i. Fresh or dried fish (e.g. kapenta) 0= No 1= Yes  Q719j. Eggs 0= No 1= Yes  Q719k. Pumpkin (squash), carrots or yellow/orange sweet potatoes, ripe mangos or pawpaws 0= No 1= Yes  Q719l. Dark green leafy vegetables such as rape, kalembula, bondwe, chibwabwa or katapa 0= No 1= Yes  Q719m Other vegetables or fruits e.g. cabbage, tomatoes, oranges, bananas, pineapple, avocados, wild fruit 0= No 1= Yes  Q719n. Cooking oil (e.g. salad), fat or foods made with oil (e.g. fried foods) 0= No 1= Yes    129  Q719o. Any sugary foods such as sweets, candies, pastries, cakes, biscuits or chocolates. 0= No 1= Yes  Q719p. Commercial infant or young child cereals (from a box, tin or jar) 0= No 1= Yes  Q719q. Any other food? 0= No 1= Yes  Q720. How many times yesterday, during the day or night, did (NAME) eat any solid or semi-solid foods (excluding breastfeeding)? ALSO PROMPT FOR SNACKS. Number of meals and snacks 77= Don’t Know  Q721. Are there any foods available to you that only adults consume and are not good for children? 0= No  End of the interview. 1= Yes  Q722. If yes, which foods are not good for children and why?  1. Food: _______________________________________  Why is not good:______________________________ ____________________________________________ 2. Food: _______________________________________  Why is it not good: ____________________________ ____________________________________________  This is the end of the interview. Thank you very much for your participation and time!  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0167049/manifest

Comment

Related Items