Assessing the impact of micronutrient interventions ...

4 downloads 37 Views 854KB Size Report
Philip Ndemwa, David Mwaniki, Erastus Muniu, and Paul- ine Andango are affiliated with the Kenya Medical Research. Institute, Nairobi. Christine L. Klotz ...
Special section

Assessing the impact of micronutrient interventions under special circumstances such as refugee camps and emergency operations

Assessing the impact of micronutrient intervention programs implemented under special circumstances— Meeting report —S. de Pee, P. Spiegel, K. Kraemer, C. Wilkinson, O. Bilukha, A. Seal, K. Macias, A. Oman, A. B. Fall, R. Yip, J. P. Pena-Rosas, K. West, S. Zlotkin, and M. W. Bloem ........................255 Effects of multimicronutrient home fortification on anemia and growth in Bhutanese refugee children —O. Bilukha, C. Howard, C. Wilkinson, S. Bamrah, and F. Husain ..........................................264 Provision of micronutrient powder in response to the Cyclone Sidr emergency in Bangladesh: Cross-sectional assessment at the end of the intervention —J. H. Rah, S. de Pee, S. Halati, M. Parveen, S. S. Mehjabeen, G. Steiger, M. W. Bloem, and K. Kraemer ..............................................277 Relationship of the availability of micronutrient powder with iron status and hemoglobin among women and children in the Kakuma Refugee Camp, Kenya —P. Ndemwa, C. L. Klotz, D. Mwaniki, K. Sun, E. Muniu, P. Andango, J. Owigar, J. H. Rah, K. Kraemer, P. B. Spiegel, M.W. Bloem, S. de Pee, and R. D. Semba..................................................................................................................286 Understanding low usage of micronutrient powder in the Kakuma Refugee Camp, Kenya: Findings from a qualitative study —S. Kodish, J. H. Rah, K. Kraemer, S. de Pee, and J. Gittelsohn ...........292

Food and Nutrition Bulletin, vol. 32, no. 3 © 2011, The United Nations University.

255

Assessing the impact of micronutrient intervention programs implemented under special circumstances— Meeting report Saskia de Pee, Paul Spiegel, Klaus Kraemer, Caroline Wilkinson, Oleg Bilukha, Andrew Seal, Kathy Macias, Allison Oman, Ahmed Baba Fall, Ray Yip, Juan Pablo Peña-Rosas, Keith West, Stanley Zlotkin, and Martin W. Bloem Abstract Introduction and Objective. The World Food Programme and the Office of the United Nations High Commissioner for Refugees organized a meeting of experts to discuss evaluation of micronutrient interventions under special circumstances, such as emergency and refugee situations. Results. Multimicronutrient interventions for groups with higher needs may include home fortification products for young children or supplements for pregnant and lactating women. The choice of preparation should be guided by target group needs, evidence of efficacy of a product or its compounds, acceptability, and costeffectiveness. Different designs can be used to assess whether an intervention has the desired impact. First, program implementation and adherence must be ascertained. Then, impact on micronutrient status can be assessed, but design options are often limited by logistic challenges, available budget, security issues, and ethical and practical issues regarding nonintervention or placebo groups. Under these conditions, a plausibility design Saskia de Pee and Martin W. Bloem are affiliated with the World Food Programme, Rome, Italy; Paul Spiegel, Caroline Wilkinson, Kathy Macias, Allison Oman, and Ahmed Baba Fall are affiliated with the Office of the United Nations High Commissioner for Refugees, Geneva, Switzerland; Klaus Kraemer is affiliated with Sight and Life, Basel, Switzerland; Oleg Bilukha is affiliated with the International Emergency and Refugee Branch, Center for Global Health, Centers for Disease Control and Prevention, Atlanta, Georgia, USA; Andrew Seal is affiliated with the Institute of Child Health, Centre for International Health and Development, University College London, London, England; Ray Yip is affiliated with the Bill and Melinda Gates Foundation, China Office, Beijing, China; Juan Pablo Peña-Rosas is affiliated with the World Health Organization, Geneva, Switzerland; Keith West is affiliated with Johns Hopkins University, Baltimore, Maryland, USA; Stanley Zlotkin is affiliated with the University of Toronto, Toronto, Ontario, Canada. Please direct queries to the corresponding author: Saskia de Pee, Nutrition and HIV/AIDS Policy Unit, Policy and Strategy Division, World Food Programme, Via Cesare Giulio Viola 68/70, 00148 Rome, Italy. Email: [email protected] or [email protected].

256

using pre- and postintervention cross-sectional surveys, a prospective cohort study, or a step-wedge design, which enrolls groups as they start receiving the intervention, should be considered. Post hoc comparison of groups with different adherence levels may also be useful. Hemoglobin is often selected as an impact indicator because it is easily measured and tends to respond to change in micronutrient status, especially iron. However, it is not a very specific indicator of micronutrient status, because it is also influenced by inflammation, parasitic infestation, physiological status (age, pregnancy), altitude, and disorders such as thalassemia and sickle cell disease. Conclusion. Given the constraints described above, replicability of impact in different contexts is key to the validation of micronutrient interventions.

Key words: Adherence, hemoglobin, impact assessment, lipid-based nutrient supplement, micronutrient powder Micronutrient (i.e., vitamin and mineral) deficiencies have long been recognized as a public health problem in many populations, particularly among high-risk groups such as young children, pregnant and lactating women, and adolescent girls. In the past, individual micronutrient deficiencies tended to be tackled separately through the use of specific supplements, such as iron and folic acid tablets for pregnant women, highdose vitamin A capsules for young children, and salt iodization for the general population. Since the late 1990s, however, the focus of prevention and treatment of micronutrient deficiencies has shifted to multiple micronutrient supplementation, for two main reasons. First, the most common cause of micronutrient deficiencies is inadequate dietary intake of micronutrients. Because foods contain combinations of nutrients, low vitamin A intake, for example, indicates that the intakes of several other micronutrients are probably low as well. Second, advanced analytical techniques and their application to biomarkers of nutrient status have confirmed that deficiencies are not limited to the

Food and Nutrition Bulletin, vol. 32, no. 3 © 2011, The United Nations University.

Impact of micronutrient intervention programs under special circumstances

most well-known micronutrients, such as iron, vitamin A, iodine, and zinc, but also include others, such as selenium, folate, and vitamins B6 and B12. Different interventions are being used to prevent and treat micronutrient deficiencies at the population level. These include supplementation; home fortification; fortification of special food products, common staples, and condiments; and improvements in dietary diversity through approaches that include homestead food production and nutrition education. As these approaches are being increasingly scaled up, it is important to monitor and evaluate their implementation and impact in order to know whether the interventions are having the desired effect and, if they are not, to identify likely causes and potential solutions. In March 2009, representatives of the World Food Programme and the United Nations High Commissioner for Refugees organized a one-day meeting of experts to discuss the evaluation of programs aimed at reducing multiple micronutrient deficiencies, particularly in emergency and refugee settings. Topics discussed included the selection of impact indicators (especially biomarkers of nutrient status), factors that must be taken into account when hemoglobin concentration is used as an impact indicator, target groups for the intervention and the evaluation, assessment of adherence to the intervention, design of the impact evaluation, and data collection. This paper reports the discussions of that meeting.

Type of programs to be evaluated The meeting discussions focused on programs that provide specific micronutrient-rich products to vulnerable populations, including micronutrient powder and low-dose (20 g/day) lipid-based nutrient supplements (LNS) to young children, micronutrient supplements to pregnant and lactating women, and special foods to specific groups such as fortified meals or snacks for school feeding or treatment of moderately malnourished children. These programs are generally implemented among all eligible individuals in a specific geographic area or setting (e.g., refugee camp, emergency-affected population, school). This has important implications for the design of the evaluation, because there is typically no control group (i.e., no one in the targeted group would receive a placebo, and everyone would be eligible to receive the intervention).

Choice of impact indicators Micronutrient deficiencies result in various clinical symptoms. Some of these symptoms, such as nightblindness and pellagra, are caused by deficiencies of specific micronutrients, whereas others, such as

257

stunting and increased morbidity, can be caused by several single or combined deficiencies. When micronutrient supplements are provided to correct a specific deficiency, impact assessments focus on the status of that micronutrient and/or the specific symptom(s) of the deficiency. In the case of high-dose vitamin A capsule distribution as a child survival intervention, for example, population-wide changes in child mortality can be monitored, whereas changes in vitamin A status among specific groups (such as schoolchildren or adult women) can be monitored over time to assess vitamin A fortification of sugar for improvement of vitamin A status. In the case of interventions with multiple micronutrients (e.g., a micronutrient powder containing 15 vitamins and minerals used for home fortification to reduce the gap between dietary intake and physiological requirements), many biological and functional outcome indicators could potentially be selected. The choice of indicators depends on several factors, including: » Indicators for which data are already available (these data may have been used to guide the decision to implement the intervention); » Specificity of the indicator (i.e., whether the indicator is likely to respond to the intervention, or whether it is also influenced by many other factors); » Responsiveness of the indicator (i.e., whether the indicator is sensitive to change in nutritional status); » Ease and cost of data collection; and » Sample size required to detect a statistically significant change of prevalence or status. Among the most monitored indicators of micronutrient status are hemoglobin concentration; indicators of iron, vitamin A, and iodine status; and indicators of infection and inflammation (e.g., alpha-1 acid glycoprotein and C-reactive protein). Among these indicators, hemoglobin concentration is the easiest to measure in the field and has been the most widely used. Functional indicators include morbidity (diarrhea, fever, acute respiratory infections), stunting, mortality, night-blindness, and developmental milestones. Most of these indicators are not specific to the deficiency of one particular micronutrient but rather reflect changes in the status of several micronutrients. Most functional indicators and indicators of micronutrient status are also affected by changes in other services or circumstances (including immunization, sanitation, and availability of clean drinking water), which are often addressed concurrently with micronutrient interventions. Programs distributing multimicronutrient preparations in order to better meet physiologic requirements (i.e., usually providing 1 RNI per dose) usually lead to several anticipated outcomes, including improved health and physical performance. For that reason, and given the criteria listed above, indicators of health and

258

nutritional status that are most commonly included in (regular) health and nutrition surveys and program evaluations are hemoglobin concentration, morbidity, and anthropometry. Additional indicators may include malaria microscopy or rapid diagnostic tests and developmental milestones. Micronutrient interventions are usually included in packages of interventions aimed at improving health and nutritional status, which makes it difficult to determine the impact of specific interventions separately (see discussion of evaluation designs below). Because hemoglobin concentration is the most widely used indicator, it was discussed in greater detail at the consultation.

Hemoglobin concentration: Considerations Anemia is widely used as a proxy for iron-deficiency anemia at the population level [1], because it is easily detected (usually by measuring hemoglobin concentration) and because 40% to 60% of anemia is due to iron deficiency [2]. However, assessment of other indicators of iron status (such as serum ferritin or serum transferrin receptor) and inflammation (to interpret ferritin values) is required to determine what proportion of anemia in a population is due to iron deficiency. Because virtually all multimicronutrient preparations include iron and other micronutrients (such as vitamins A and C, folic acid, and vitamin B12) that could improve hemoglobin, hemoglobin has been the indicator of choice for assessing the programmatic impact of multimicronutrient interventions. Other biochemical indicators, such as indicators of iron or vitamin A status or infection, were included in only a very few cases, mainly because of cost and the increasing complexity of sample collection, handling, storage, and analysis. However, anemia is also affected by non-nutritional causes, such as parasitic infestation, malaria, HIV/AIDS infection, thalassemia, sickle cell disease, and physiological factors such as age and pregnancy. Some of these factors can be monitored, whereas others are more difficult to detect. Therefore, although the reasons for widespread use of hemoglobin as an indicator of change of micronutrient status are clear, its constraints must be well understood in order to interpret assessment findings appropriately.

S. de Pee et al.

Standardization of methods of blood collection (including fingerprick) and measurement is essential. Because some types of HemoCue cuvettes are hygroscopic, keeping a cuvette container open in humid climates may affect the results. In that case, it is very important to minimize the time during which containers are open and to discard cuvettes within approximately 2 weeks of breaking the seal of any individual container. Cutoff

The meeting participants agreed that the current hemoglobin cutoffs for anemia (table 1) are appropriate but noted that a cutoff of 100 g/L would be better for children aged 3 to 9 months [4]. Whereas the latter is relevant for individual measurements or when only measuring children in this specific age range, it is most practical to use the cutoff of 110 g/L for the target group of children aged 6 to 59 months. The experts also agreed that the cutoffs indicating mild, moderate, or severe levels of anemia should be maintained and used in reporting of results. Altitude

Hemoglobin concentration is higher at higher altitudes because of the lower oxygen concentration in the air. Therefore, hemoglobin cutoffs for anemia are higher at higher altitudes (table 2). A correction for altitude should be made when hemoglobin or anemia levels at different altitudes are compared. Seasonality

Hemoglobin concentration decreases because of infection, chronic inflammation, or malaria. Therefore, it could respond to seasonal changes, especially related to morbidity patterns. Because of this, it is best when repeated assessments are conducted in the same season, i.e., at 12-month intervals. Age

Because of rapid expansion of circulating blood volume, iron needs in children are highest before 12 TABLE 1. Cutoffs for hemoglobin concentration that indicate anemia, by target group

Assessment

Target group

Hemoglobin concentration can be most easily measured with the HemoCue device, which is portable, uses a rechargeable battery, and requires just one drop of blood (often obtained by fingerprick). Other methods include collection of blood on filter paper (and dissolution in the laboratory) and analysis using Drabkin’s solution, which requires a spectrophotometer [3].

Children 6–59 mo Children 5–11 yr Children 12–14 yr Nonpregnant women ≥ 15 yr Pregnant women Men ≥ 15 yr Source: WHO/UNICEF/UNU [1].

Cutoff (g/L) 110 115 120 120 110 130

Impact of micronutrient intervention programs under special circumstances

TABLE 2. Normal increase in hemoglobin concentration related to long-term altitude exposure Altitude (m)

Increase (g/L)

< 1,000 1,000 1,500 2,000 2,500 3,000 3,500 4,000

0 2 5 8 13 19 27 35

4,500

45

Source: WHO/UNICEF/UNU [1].

months of age and then gradually decrease, while hemoglobin increases. It is therefore important that interventions specifically target the youngest children and that when following hemoglobin in cohorts of children over time, the natural increase of hemoglobin with age is accounted for when estimating the impact of the intervention. Physiological status

Pregnant women have the highest iron needs and, as a result, the highest prevalence of anemia among adults in iron-deficient populations. A high prevalence of anemia among nonpregnant women and young children is indicative of an even higher prevalence among pregnant women. After delivery, hemoglobin returns to prepregnancy levels, as long as pregnancy did not induce or worsen iron deficiency. The return to prepregnancy values is aided by lactational amenorrhea, which conserves iron. When anemia prevalence among nonpregnant women is reported, both lactating and nonlactating women can be included (i.e., they do not need to be considered separate groups). Anemia in pregnant women should be reported separately. Postmenarcheal adolescent girls are also at risk for iron deficiency. After menopause, women’s iron needs are comparable to those of adult men, who have the lowest iron needs. Because of their low iron requirement, anemia among adult men is rarely due to iron deficiency, but rather to non-nutritional causes such as thalassemia, parasitic infestation, or other pathology. The meeting participants liked Yip’s recommendation to measure anemia prevalence among a sample of adult men to get an indication of whether non-nutritional causes of anemia are prevalent in a population.

Target groups for the intervention and the evaluation As discussed above, the groups with the highest needs for iron and other micronutrients are usually infants

259

and young children, pregnant women, and women and girls of reproductive age. However, it is important to note that some deficiency diseases, such as pellagra (vitamin B3 deficiency), preferentially affect adults, including men. Vulnerable groups should be targeted with micronutrient interventions in populations where deficiencies are widely prevalent. In order to specifically increase micronutrient intake among these groups, the most appropriate interventions are those that provide supplements in the form of micronutrient powder or a low-dose micronutrient-rich spread (approximately 20 g/day) for home fortification (most appropriate for young children), capsules or tablets (for older age groups), or special food preparations such as micronutrient-rich snacks or drinks. Although fortification of general foods (such as staples and condiments) and increasing the intake of animal-source foods, fruits, and vegetables to the extent possible will increase the intake of micronutrients, additional sources may be required in order to meet the high micronutrient needs of vulnerable groups. A high anemia prevalence indicates a very likely gap between intake and requirements for iron as well as other micronutrients. The obvious choice for assessing impact is to collect data among the targeted groups, for example, by following a cohort. However, assessing and interpreting change over time is somewhat tricky among certain groups, in particular among pregnant women and children under 2 years of age. Pregnant women

Tracking hemoglobin changes among a cohort of pregnant women is challenging, for a number of reasons. First, pregnant women are likely to start interventions at different times during pregnancy, depending on when they first seek antenatal care. An intervention during pregnancy will usually last only 4 to 6 months (depending on timing of the first antenatal visit), and a certain number of supplements needs to be consumed before an impact can be achieved. Physiological changes, such as hemodilution, affect hemoglobin and the serum concentration of micronutrients. Furthermore, finding enough pregnant women in a narrow starting range of gestational age is also very challenging when operating in a programmatic setting. Thus, tracking hemoglobin changes among pregnant women is very difficult, except in carefully controlled circumstances and starting with early detection of pregnancy. The impact of antenatal supplementation may be best assessed by comparing lactating women who have been exposed to the intervention during pregnancy with women with a child of the same age who were not exposed during their pregnancy. This could be carried out by a cross-sectional comparison (i.e., between a sample from the intervention area and one from a

260

nonintervention area) or a comparison of a sample of lactating women measured before the program started and another sample from the same population measured after the program has been in place for a certain length of time. Children under 5 years of age

The challenge of measuring hemoglobin among children under 5 years old is ageing, which causes a physiologic increase of hemoglobin that must be accounted for when interpreting hemoglobin change within a cohort. This is especially important when the follow-up period is longer than a couple of months. Taking the above points into consideration, the meeting participants discussed possible methods for designing program evaluations. The results of their discussions are presented below.

Assessing adherence In order to be able to interpret the findings of an impact assessment, it is very important to monitor the actual implementation of the program: » Have the eligible individuals received the intervention? » Have they understood why they received the product? » Have they accepted the product? » Have they used the product in the recommended way? (Was the product added to the food of eligible individuals? Was it added to food that was ready for consumption? Did the participants adhere to the recommended dosing frequency?) The findings of the impact assessment can be interpreted only after it has been ascertained that the product was consumed as intended. However, obtaining reliable information on adherence is often difficult. To verify answers obtained from beneficiaries or their caregivers, additional information can be sought by asking to see the supplement container and comparing the actual and expected numbers of remaining sachets. Information can also be collected from beneficiaries’ consumption calendars for programs in which such calendars are provided. Researchers who are collecting and interpreting adherence data must also take into account the supplement delivery schedule and instructions for use. If 30 single doses are provided at 2-month intervals (often referred to as “flexible dosing”), for example, some individuals may finish all doses in the first month, whereas others may have finished them only by the end of the second month (adherence in both cases would be considered good). In this case, adherence assessment would be best conducted at the end of the second month to assess whether all doses had been finished.

S. de Pee et al.

Design of the impact evaluation Various impact evaluation designs can be used to determine whether a particular intervention that has been proven efficacious—or can be assumed to be efficacious based on supplement composition*—indeed has had the desired impact on the target population. As mentioned above, it is difficult to ascertain impact when an intervention is provided to every eligible person in the target population because of the following factors**: » Lack of a control group of randomly selected individuals, comparable to the individuals receiving the intervention, who would receive a placebo; » The fact that other factors that could affect the outcome measure (such as infections, seasonality, change in age or physiological status, and concurrent implementation of other health and nutritional interventions) cannot be controlled for and that the exposure to such factors thus needs to be monitored and their impact estimated rather than measured because of the lack of a placebo-receiving control group; » Limited control over actual consumption of the micronutrient-rich commodity (distribution can be monitored, but adherence relies on self-reporting). The meeting participants discussed possible evaluation designs for use in situations such as emergency responses, in which it is not feasible to include a randomly selected control group and/or conduct both baseline and follow-up or endline measurements. These designs are presented below. The participants also referred to the concept of distinguishing adequacy, plausibility, and probability designs for evaluating the impact of public health programs, as proposed by Habicht and colleagues [5]. Cohort tracking or cross-sectional assessments

In many cases, it would be possible to follow a cohort of subjects from before the start of the program or intervention (i.e., baseline) to follow-up or endline * As micronutrient formulation of different micronutrient powders and LNS varies according to situation, it is difficult to obtain efficacy data on all different formulations. However, each product should be formulated using ingredients (or chemical forms of specific nutrients) that have been shown to be efficacious, if not necessarily in the specific selected combinations. ** An exceptional case would be a situation in which a new product was provided in lieu of another product and both products were comparable with regard to nutrient content and formulation but potentially different with regard to acceptability. In this case, one group could continue to receive the already used product while the other group could receive the new product. The impact evaluation would then assess whether the impacts of the two products were equal rather than different, and any difference would be related to acceptance and use of the products rather than to their composition.

Impact of micronutrient intervention programs under special circumstances

measurements and to conduct one or more visits to monitor adherence between the baseline and endline. Tracking a cohort is a good option, because assessing change within the same subjects is often more sensitive than assessing change among two groups of subjects measured at different points in time. However, as it is important that data from the cohort represent changes in the rest of the intervention population, it is important to avoid too frequent contact with the cohort that could bias their behavior and/or attitudes toward the intervention. Where possible, data from the cohort must be corroborated with data from cross-sectional surveys among the intervention population in order to verify that the observations among the cohort reflect changes among the larger populations. Those cross-sectional surveys could include indicators of micronutrient status as well as adherence. Tracking of a cohort is not a good method when an intervention is evaluated over a period of time that is relatively long compared with the period during which individuals are eligible for the intervention. In such a case, it is better to use repeated cross-sectional surveys. For logistics and monitoring considerations, the samples could be selected each time from the same clusters if cluster design is used. As with the tracking of a cohort, it would be important to ensure that these clusters are treated more or less the same as the nonsample clusters to ensure comparability with the rest of the population exposed to the intervention. In some cases, it may not be possible to conduct a baseline measurement, particularly in the context of emergency programming. Under such circumstances, researchers may consider comparison of cross-sectional data collected from intervention and nonintervention areas, possibly enhanced by inclusion of a nontargeted group (see below). Comparison group from outside the program area

When evaluating the impact of programs implemented in a certain geographic area, the same target group from nonprogram areas can serve as a comparison group. Such data can be collected purposively (i.e., by the same data collection teams using the same tools and at the same time), or secondary data from sources such as a Demographic and Health Survey can be used. Data from the comparison and intervention groups should preferably be available for the same moment in time. However, it may be difficult to ascertain a relatively limited impact by comparing data from intervention and comparison groups if members of the comparison group have been excluded from the intervention for a specific reason. Comparison group members often have less need for the intervention, which may mean that they are likely to have a somewhat better nutritional status than members of the intervention group. Likewise, selecting a suitable comparison group

261

is usually not possible in refugee situations, because nearby nonrefugee populations do not receive the same food assistance and hygiene and preventive health services. Another way of choosing a more suitable comparison group is by phasing-in of the intervention. Step-wedge design or phasing-in of the intervention

Some groups within a target population may receive an intervention earlier than others if there are limitations on the capacity of introducing such an intervention to the whole target population. This may include limited supplies, technical expertise, or personnel or other programmatic limitations. Individuals who receive an intervention later can serve as a comparison group for those who received it earlier. Dose–response relationship among subjects of the intervention group

With any program, there will be a gradient of adherence. Therefore, those who do not adhere to the intervention well can serve as the comparison group for those who do adhere well. In order for such a comparison to have enough statistical power to detect a difference, there should be enough variability in adherence among the measured subjects and adequate numbers of individuals with both low and high adherence to make a valid comparison. In order to collect good adherence data, there should also be regular (e.g., every two months) household visits. However, in order to ensure as much similarity to the intervention population as possible, these visits should not be too frequent. A nontargeted group as a comparison group

When one target group, for example, children 6 to 59 months of age, receives the intervention, another target group, for example, adolescent girls or schoolage children, could serve as a comparison group. With this comparison, researchers would compare changes in levels of anemia or hemoglobin over time rather than absolute levels of anemia or hemoglobin in the targeted and nontargeted groups. The hypothesis would be that the group who received the intervention would show a different change over time (improvement or no change) compared with the nontargeted group (no change or deterioration). The “nontargeted” comparison group should be chosen so that the etiology of anemia is not very different between the groups. Thus, for young children, postmenarcheal adolescent girls who have not received iron-rich foods or supplements would be a good comparison group. Including a nontargeted comparison group can also be considered when only a postintervention comparison is conducted between an intervention and a

262

nonintervention area, for example, in emergencies when baseline measurements have not been conducted. This would be in addition to comparing the intervention area’s group of eligible subjects and the nonintervention area’s group of subjects within the same target group [6].

Specific data collection issues Most data collection issues discussed during the meeting concerned hemoglobin assessment and adherence (see above). The following additional topics were mentioned briefly. Sampling

Where possible, sampling should be conducted by simple or systematic random sampling rather than cluster sampling. Ethical clearance

No specific ethical permission is required for crosssectional data collection that is part of routine program monitoring and does not collect identifier information (name and address). However, informed consent (according to Helsinki Declaration standards) must be obtained from all subjects before starting data collection, and ethical permission must be obtained for all nonroutine data collection, including cohort studies and surveys, as well as surveys that collect identifier data. Consistency and quality of data collection

In order to collect good-quality data and to be able to detect subtle but meaningful changes, it is important that the data collection methods and training of enumerators are standardized, and all data should preferably be collected by the same institution or team over time.

S. de Pee et al.

Conclusions Micronutrient deficiencies, as indicated, for example, by a high anemia prevalence, often trigger implementation of micronutrient interventions, such as introduction of home fortification for young children. Such interventions are often conducted in combination with other public health programs. The choice of preparation should be guided by target group needs, evidence of efficacy of a product or its compounds, acceptability, and cost-effectiveness. In order to determine whether the intervention achieves its intended goal, i.e., a reduction of the prevalence of micronutrient deficiencies and their consequences, the micronutrient intervention and other related interventions must be monitored and evaluated. Monitoring focuses on program implementation and on ascertaining that the intervention is accepted and adhered to by the beneficiaries. Impact evaluation requires a choice of impact indicator(s) and appropriate evaluation design. Hemoglobin concentration is often selected as an impact indicator because it is easy to measure and usually responds to changes in micronutrient status, especially of iron. However, it is not a very specific indicator of iron deficiency or the status of other micronutrients, because it can be influenced by infection and inflammation, parasitic infestation, physiological status (e.g., age and pregnancy), and blood disorders such as thalassemia and sickle cell disease. Thus, the possible impact of a multimicronutrient intervention on hemoglobin varies among populations, and an impact (or lack thereof) on hemoglobin concentration does not provide ultimate proof of impact or absence of impact of the micronutrient intervention. Inclusion of additional indicators, such as those of growth and morbidity, as well as measuring the status of other micronutrients, can provide further insights. For special situations (such as emergencies and refugee settings) in which all eligible subjects receive the intervention and a suitable comparison group or area may not be available, alternative evaluation designs can be considered. These include assessment of a cohort over time, repeated cross-sectional surveys, inclusion of a nontargeted group to compare change between targeted and nontargeted groups, or comparison with a nonintervention group, for example, one that is exposed to the intervention at a later stage (step-wedge design) or is a subgroup with poor adherence to the intervention.

References 1. World Health Organization/UNICEF/United Nations University. Iron deficiency anemia: assessment, prevention and control. Geneva: WHO, 2001. 2. Black RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzafi M, Mathers C, Rivera J, for the Maternal and Child Undernutrition Study Group. Maternal and child

undernutrition: global and regional exposures and health consequences. Lancet 2008;371:243–60. 3. Sari M, de Pee S, Martini E, Herman S, Sugiatmi, Bloem MW, Yip R. Estimating the prevalence of anaemia: a comparison of three methods. Bull World Health Organ 2001;79:506–11.

Impact of micronutrient intervention programs under special circumstances

4. Domellöf M, Cohen RJ, Dewey KG, Hernell O, Rivera LL, Lönnerdal B. Iron supplementation of breast-fed Honduran and Swedish infants from 4 to 9 months of age. J Pediatr 2001;138:679–87. 5. Habicht JP, Victora CG, Vaughan JP. Evaluation designs for adequacy, plausibility and probability of public health programme performance and impact. Int J

263

Epidemiol 1999;28:10-8. 6. Rah JH, de Pee S, Halati S, Parveen M, Mehjabeen SS, Steiger G, Bloem MW, Kraemer K. Provision of micronutrient powder in response to the Cyclone Sidr emergency in Bangladesh: cross-sectional assessment at the end of the intervention. Food Nutr Bull 2011;32:276–84.

Effects of multimicronutrient home fortification on anemia and growth in Bhutanese refugee children

Oleg Bilukha, Christopher Howard, Caroline Wilkinson, Sapna Bamrah, and Farah Husain Abstract Background. Anemia remains a significant public health problem in refugee settings. Home fortification with micronutrient powders has been proposed as a feasible option to alleviate micronutrient deficiencies; its efficacy in reducing anemia in children aged 6 to 24 months has been demonstrated in several trials. Objective. To evaluate the effectiveness of a large-scale micronutrient powder distribution program in reducing anemia prevalence and promoting growth in refugee children aged 6 to 59 months. Methods. Four representative cross-sectional surveys were conducted 13 months before and 7, 14, and 26 months after initiation of the supplementation program. Data collected on children aged 6 to 59 months included hemoglobin concentration, anthropometric indicators, morbidity, feeding practices, and information on the micronutrient distribution program. The study had a pre–post design with no control group. Results. The overall prevalence of anemia in children did not change significantly between baseline (43.3%) and endpoint (40.2%). The prevalence of moderate anemia decreased over the same period from 18.9% to 14.4% (p < .05). The levels of severe anemia were negligible (< 1%) in all surveys. The prevalence of stunting decreased significantly from 39.2% at baseline to 23.4% at endpoint (p < .001), a relative decrease of 40%. Reported coverage, use, and acceptance of micronutrient supplements remained consistently high throughout the study. Conclusions. In the absence of a control group, changes in key outcomes should be interpreted with Oleg Bilukha, Christopher Howard, Sapna Bamrah, and Farah Husain are affiliated with the Centers for Disease Control and Prevention, Atlanta, Georgia, USA; Caroline Wilkinson is affiliated with the Office of the United Nations High Commissioner for Refugees, Geneva. Please direct queries to the corresponding author: Oleg Bilukha, CDC, 4770 Buford Hwy, MS F-60, Atlanta, GA 30341, USA; e-mail: [email protected].

264

caution. The minor effect on hemoglobin status requires further investigation of underlying causes of anemia in this population. The large positive effect on linear growth may be a significant benefit of supplementation if confirmed by future studies.

Key words: Anemia, effectiveness, growth, home fortification, micronutrient powder, morbidity

Introduction Anemia remains a serious public health problem, with close to 50% of preschoolchildren worldwide affected [1]. An estimated 40% to 60% of anemia is due to iron deficiency [2], with other causes including other nutritional deficiencies, inflammation, malaria, parasite infestation, and hemoglobinopathies, among other factors. Iron-deficiency anemia adversely affects cognitive development, physical growth, immune status, morbidity from infectious disease, and work performance, among other outcomes [3, 4]. Previous studies have consistently shown high prevalence rates of anemia in refugee settings [5–7]. This is attributed in part to poor access to foods rich in iron and other essential micronutrients because of refugees’ reliance on restricted food rations and physical and financial constraints on access to food markets. The cerealbased diets commonly provided to refugees are a poor source of bioavailable iron. Substandard infant feeding practices and frequent illness are among other factors contributing to anemia in refugee settings. Home fortification with micronutrient powder has been proposed in the past decade as one of the feasible options to alleviate anemia at the population level in developing countries [8]. Several efficacy trials have shown that iron-containing micronutrient powder administered daily [9–14], weekly [15], or according to flexible regimens [16] is effective in treating moderate anemia (hemoglobin level, 70 to 100 g/L)

Food and Nutrition Bulletin, vol. 32, no. 3 © 2011, The United Nations University.

265

Micronutrient intervention and anemia in refugee children

and increasing mean hemoglobin levels in infants and young children aged 6 to 24 months in developingcountry settings with high anemia prevalence. On the other hand, micronutrient powder had no effect on hemoglobin or iron status in mildly anemic and nonanemic children [17]. The micronutrient content of home fortification powders used in most recent studies included iron, zinc, and folate, as well as vitamins A, C [11, 15, 16, 18], and D [12–14]. The objective of our study was to evaluate the effectiveness of a large-scale micronutrient powder distribution program in reducing anemia and to monitor morbidity and growth in refugee children aged 6 to 59 months. Setting and programmatic context

In 1991, approximately 100,000 Bhutanese, mostly of Nepali origin, fled Bhutan because of ethnic persecution and were settled in seven refugee camps in two adjacent districts of southeastern Nepal. The refugee population was relatively stable between 1993 and 2008. The program of resettlement to third countries started in 2008. By mid-2010, of approximately 112,000 refugees registered in 2008, 32,000 had been resettled and about 80,000 remained in the camps. All refugees in the camps received a food ration that provided 2,100 kcal/day and consisted of rice (400 g/day), dried beans (40 g/day), lentils (20 g/day), sugar (20 g/day), iodized salt (8 g/day), vitamins A- and D-fortified vegetable oil (25 g/day), and fortified wheat–soy blend (Unilito) (35 g/day). Children aged 6 to 59 months with moderate acute malnutrition were enrolled in a supplementary feeding program and received 200 g of Unilito mixed with 20 g of oil daily. All refugees had free access to health services, including immunizations. Children aged 6 to 59 months received vitamin A supplements (200,000 IU) twice a year, and children aged 12 to 59 months were dewormed with albendazole following the same twice a year schedule. Through the home gardening program implemented on a wide scale in all camps, refugee families were encouraged, educated, and supported to grow fresh fruits and vegetables for personal consumption in small kitchen gardens. The micronutrient powder intervention

The micronutrient powder distribution program was initiated in March 2008, based on the results of the nutrition survey conducted in 2007 that showed a 43% prevalence of anemia among children aged 6 to 59 months [19]. The micronutrient powder used in the Bhutanese refugee camps, Vita-Mix-It, contained 16 vitamins and minerals (table 1) and was based on the Recommended Nutrient Intake (RNI) for children 1 to 3 years of age [20, 21]. Compared with RNIs, the VitaMix-It contained reduced amounts of iodine (30 µg

versus 90 µg RNI) and vitamin A (100 µg vs. 400 µg RNI), because the refugee food ration contains iodized salt as well as oil and Unilito fortified with vitamin A. The content of vitamin C (30 mg initially) was doubled to 60 mg starting in April 2009 to further enhance iron absorption. Refugee children aged 6 to 59 months received one Vita-Mix-It sachet every 2 days, which provided an average of 50% RNI per day for most vitamins and minerals. This regimen was based on the fact that children in this population consumed a diet that contained some micronutrients from fortified and natural foods, as well as previous evidence that flexible administration of 60 sachets over 3 to 4 months was efficacious in reducing anemia in children aged 6 to 24 months [16]. The distribution was conducted monthly, with each child receiving approximately 15 sachets for each 30-day period. The distribution was linked to monthly growth-monitoring activity, and refugee health and nutrition workers actively followed up defaulters who did not receive their monthly micronutrient powder distribution. Children were enrolled in the program as soon as they reached the age of 6 months, and mothers received education on the purposes, benefits, and proper use and storage of Vita-Mix-It at enrollment. The mothers were instructed to mix the contents of one full sachet with the child’s meal after it had been cooked. Instructions on use and storage in the Nepali language were included on each individual sachet. Each month, health and nutrition workers conducted home visits to approximately 50% of enrolled children on a rotation basis to monitor use and assess possible problems. From February 2009, education for mothers also included messages on appropriate introduction and use of complementary foods as well as messages to reduce tea consumption among children. Individual TABLE 1. Nutrient composition of Vita-Mix-It Vitamin or mineral Vitamin A—μg Vitamin D—μg Vitamin E—mg Vitamin C—mg Thiamine (vitamin B1)—mg Riboflavin (vitamin B2)—mg Niacin (vitamin B3)—mg Pyridoxine (vitamin B6)—mg Cobalamin (vitamin B12)—μg Folic acid—μg Iron—mg Zinc—mg Copper—mg Selenium—μg Iodine—μg Vitamin K—μg

Amount per 1-g sachet 100.0 5.0 5.0 60.0 0.5 0.5 6.0 0.5 0.9 150.0 10.0 4.1 0.34 17.0 3.0 30.0

266

Vita-Mix-It distribution cards were introduced in mid-2009 to better track coverage. These cards were issued to each child enrolled in the program, and a record indicating the distribution date and the number of sachets provided was made on the card at each distribution.

Methods Study design and sample size

We conducted a series of population-representative, cross-sectional surveys to assess changes in anemia and anthropometric indicators among refugee children aged 6 to 59 months. These surveys were a part of the routine health and nutrition monitoring conducted in the camps, and from 2008 onwards they were also specifically designed to evaluate the micronutrient powder distribution program. Because Vita-Mix-It intervention was introduced in all camps simultaneously, a control group was not available and the evaluation followed a pre–post design with no controls. The first survey was conducted in January and February 2007, approximately 13 months before the Vita-Mix-It intervention was introduced, and served as baseline. The subsequent surveys were conducted in October 2008, May 2009, and May 2010 and provided data at 7, 14, and 26 months after the initiation of Vita-Mix-It distribution. The sample size for the 2007 survey was calculated based on 8% expected prevalence of acute malnutrition, ± 2.5% required precision, and expected 5% nonresponse rate, resulting in a sample size of 480 children. Because updated census data on children were not available in 2007, we used the lists of all refugee households provided by the Office of the United Nations High Commissioner for Refugees. Assuming on average 0.4 children aged 6 to 59 months per household, we randomly selected 1,189 households, and all eligible children residing in these households were included in the survey. The sample size in the 2008–10 surveys was calculated to statistically detect an expected 20% relative decrease in the prevalence of anemia compared with the previous year’s prevalence, with .05 alpha and 80% power. The expected nonresponse rate was set at 15% in 2008 and increased to 25% for 2009 and 2010 because of ongoing resettlement and the fact that many registered refugee families resided outside the camps. The number of children selected for the 2008, 2009, and 2010 surveys was 575, 675, and 675, respectively. In the 2008–10 surveys, we randomly sampled children aged 6 to 59 months using the updated individual refugee registration data from the Office of the United Nations High Commissioner for Refugees. Children in the database were identified by birth date, sex, and residence address. To select survey samples, we used lists of children who were aged 6 to 59 months at the

O. Bilukha et al.

time of the surveys. All surveys were representative of all seven camps, with the number of children selected from each camp proportional to the size of the camp. Measurements

The information on children and their mothers was collected by trained health and nutrition workers in interviews with the child’s mother or caregiver. We collected data on basic demographics, breastfeeding and complementary feeding practices, recent illness and care-seeking behaviors, and coverage of vitamin A distribution and deworming programs. Acute respiratory infection was defined as fever and cough or difficulty breathing in the preceding 14 days. Diarrhea was defined as three or more loose watery stools in a 24-hour period during the preceding 14 days. Surveys in 2008–10 also included questions about the VitaMix-It distribution program (distribution coverage, current intake, perceived negative health effects, perceived changes to food, etc.). In addition, the survey in 2010 collected data on the availability of Vita-Mix-It distribution cards and asked closed-ended questions on perceived changes in the child’s health, appetite, and activity level. Hemoglobin was measured in capillary blood by trained technicians with a HemoCue B-Hemoglobin photometer (HemoCue Hb 201+ Analyzer) following a standard procedure [22]. Hemoglobin in children aged 6 to 59 months was measured in all four surveys. In order to monitor hemoglobin status in refugee women, in the 2007 and 2009 surveys we also measured hemoglobin in the mothers of the children included in the survey. The mothers did not receive micronutrient powder supplements through this distribution program. Their pregnancy status was ascertained by self-report. Anemia in children aged 6 to 59 months was classified according to World Health Organization (WHO) definitions [23] as severe (hemoglobin < 70 g/L), moderate (70 g/L ≤ hemoglobin < 100 g/L), or mild (100 g/L ≤ hemoglobin < 110 g/L). Since many previous efficacy studies used hemoglobin ≤ 100 g/L as a recovery endpoint [9–11, 13, 17], we also analyzed the percentage of children with hemoglobin < 100 g/L. In nonpregnant women of reproductive age, anemia was classified as severe (hemoglobin < 80 g/L), moderate (80 g/L ≤ hemoglobin < 110 g/L), or mild (110 g/L ≤ hemoglobin < 120 g/L) [23]. Weight and height in children aged 6 to 59 months were measured by trained nutrition staff following standard procedures [24]. Weight was measured with a digital Seca 881 U scale, accurate to 0.1 kg, and height (or recumbent length for children under 2 years of age) was measured with a Shorr Infant-Child Height Board, accurate to 1 mm. Children’s age was ascertained from the Road-to-Health cards that are distributed to refugee children at birth and are used to record key health

267

Micronutrient intervention and anemia in refugee children

Results

and nutrition information (immunizations, vitamin A distribution, growth monitoring data, etc.). The presence of bilateral pitting edema was ascertained at the time of anthropometric measurements. Standardized z-scores based on 2006 WHO Growth Standards [25] were generated with the use of Epi Info/ENA software [26]. Acute malnutrition (based on weight-for-height z-score and the presence of edema), stunting (based on height-for-age z-score), and underweight (based on weight-for-age z-score) were defined and classified as moderate (–3 ≤ z-score < –2) and severe (z-score < –3) according to WHO definitions [25]. Extreme outliers, defined on the basis of 2006 WHO cutoffs (height-forage z-score < –6 or > + 6, weight-for-age z-score < –6 or > + 5, weight-for-height z-score < –5 or > + 5) were excluded from the analysis.

Survey samples

In 2007, 497 eligible children were present at the time of the survey in 1,189 selected households and were included in the study. In the 2008, 2009, and 2010 surveys, 502 (of 575 selected), 568 (of 675 selected) and 569 (of 675 selected) children, respectively, were present at the time of the survey and were included in the study. Absent children had resettled, had died, or were living outside the camps. The absentee rates were 12.7%, 15.9%, and 15.7% in 2008, 2009, and 2010, respectively, which were lower than anticipated during sample size calculations. None of the mothers or caretakers of children who were present declined permission for their children to participate in any of the surveys. There were no significant differences in the distributions of age, sex, or camp of residence between the initially selected and the final samples in any of the surveys. Two children in the 2009 survey had heightfor-age z-scores lower than –6 and were excluded from analysis of stunting. None of the children in the four surveys had weight-for-height or weight-for-age z-scores out of range. The age and sex distributions of the survey samples are presented in table 2.

Statistical analysis

Data were analyzed with JMP, version 8.0.2 (SAS Institute), and SPSS, version 17.0. P values less than .05 were considered to indicate statistical significance. All statistical tests were two-tailed. Analysis of outcome variables was based on intention-to-treat, regardless of reported intake of Vita-Mix-It supplements. Comparisons of means were computed using Tukey’s adjustment for multiple pairwise comparisons. Differences in categorical variables ware assessed by chi-square tests.

Anemia

Overall, the prevalence of anemia in children remained relatively stable at around 40%. Anemia prevalence was significantly lower than in 2007 (43.3%) only in 2009 (35.9%), whereas in 2008 and 2010 the prevalence (43.6% and 40.2%, respectively) did not differ significantly from the 2007 baseline (table 3). Only three children in the four surveys (two in 2008 and one in 2009) had severe anemia. Therefore, almost all children with hemoglobin < 100 g/L were moderately anemic. The prevalence of moderate anemia decreased from 18.9% in 2007 to 14.4% in 2010. The mean hemoglobin level was highest in 2009 (112.9 g/L) and lowest in 2008 and

Ethics

Survey protocols were exempted from review by the Human Subjects Review Board of the Centers for Disease Control and Prevention. The surveys were determined to constitute routine program monitoring activities, where collected data are used for disease control program or policy purposes. Mothers or caregivers were informed about the purpose and procedures of the study, and oral informed consent was obtained. Personal identifiers (names and addresses of children and their mothers) were not collected. TABLE 2. Survey sample characteristics

Time of survey Characteristic

Jan 2007 (n = 497)

Oct 2008 (n = 502)

May 2009 (n = 568)

May 2010 (n = 569)

Male sex—% Mean age—mo

48.7 31.7

53.0 31.7

47.8 32.1

52.5 32.3

Age (mo)—% of survey sample 6–11 12–23 24–35 36–47 48–59

10.4 23.5 23.9 19.9 22.1

9.2 22.7 26.1 19.9 22.1

11.4 24.8 21.7 21.5 20.6

10.7 23.6 20.9 24.4 20.4

268

O. Bilukha et al.

TABLE 3. Anemia prevalence and mean hemoglobin levels, by age group and time of surveya Time of survey

Variable

May 2010 May 2009 Oct 2008 Jan 2007 (total n = 497; 6–23 (total n = 502; 6–23 (total n = 568; 6–23 (total n = 569; 6–23 mo, n = 169; 24–59 mo, n = 160; 24–59 mo, n = 206; 24–59 mo, n = 195; 24–59 mo, n = 374) mo, n = 362) mo, n = 342) mo, n = 328)

Anemia prevalence—% (95% CI) 6–59 mo Mild (Hb 100.0–109.9 g/L) Moderate (Hb 70.0–99.9 g/L) Severe (Hb < 70 g/L) Total (Hb < 110 g/L)

24.3 (20.7–28.4)A 18.9 (15.6–22.7)A 0.0 43.3 (36.3–51.1)A

25.1 (21.4–29.2)A 21.1 (17.9–24.8)A 18.1 (14.9–21.8)A,B 14.6 (11.9–17.8)A,B 0.4 (0.1–1.6) 0.2 (0.0–1.1) 43.6 (39.3–48.1)A 35.9 (32.0–40.0)B

25.8 (22.3–29.7)A 14.4 (11.7–17.6)B 0.0 40.2 (34.0–47.3)A,B

36.9 (29.4–44.9)A 70.0 (62.3–77.0)A,B

29.1 (23.0–35.8)A 59.7 (52.7–66.5)C

27.7 (21.5–34.5)A 61.5 (54.3–68.4)B,C

24–59 mo Moderate or severe (Hb < 100 g/L) 11.3 (8.2–15.3)A 9.9 (7.1–13.7)A,B 6.6 (4.4–9.8)B Total (Hb < 110 g/L) 28.7 (23.9–33.9)A,B 31.3 (26.5–36.5)A 22.4 (18.3–27.1)B

7.5 (5.1–10.8)A,B 29.1 (24.6–34.1)A

6–23 mo Moderate or severe (Hb < 100 g/L) 33.7 (26.6–41.4)A Total (Hb < 110 g/L) 71.6 (64.2–78.3)A

Mean ± SD Hb (g/L) Total (6–59 mo) 6–23 mo 24–59 mo

111.6 ± 14.1A,B 103.6 ± 11.4A 115.7 ± 13.6A,B

111.0 ± 12.7A,B 103.6 ± 11.3A 114.4 ± 11.9B,C

112.9 ± 12.1A 105.8 ± 11.3A 116.9 ± 10.7A

110.7 ± 10.8B 105.3 ± 10.9A 113.5 ± 9.7C

Hb, hemoglobin a. Values in the same row not sharing the same capital superscript letter are significantly different (p < .05).

Growth

The prevalence of global acute malnutrition increased significantly from 4.2% in 2007 to 9.2% in 2008 and did not show significant changes thereafter (table 4). The

mean weight-for-height z-score improved significantly from –0.85 in 2008 to –0.65 in 2010 but still remained well below the –0.49 baseline level observed in 2007. There were no children with bilateral pitting edema in any of the surveys. The prevalence of underweight increased significantly from 20.9% in 2007 to 28.1% in 2008 and then returned to 2007 levels in 2009 and 2010 (table 4). The prevalence of stunting decreased every year from 2007 to 2010. It dropped significantly from 39.2% in 2007 to 23.4% in 2010 (p < 0.001), a relative decrease of 40%. The prevalence of severe stunting decreased from 8.0% in 2007 to 3.2% in 2010, a relative decrease 80

Prevalence of anemia (%)

2010 (111.0 and 110.7 g/L, respectively). The prevalence of anemia was much higher in younger children (aged 6 to 23 months) than in older age groups (fig. 1). Among the younger age group, anemia prevalence significantly decreased from 71.6% in 2007 to 61.5% in 2010 (table 3). The mean hemoglobin level in this age group increased from 103.6 g/L in 2007 to 105.3 g/L in 2010, but this difference did not reach statistical significance. Among older children (aged 24 to 59 months), the prevalence of anemia was around 30%, except for 2009 when it was significantly lower (22.4%). The proportion of children with hemoglobin < 100 g/L in this age group was approximately 7% in 2009 and 2010. The mean hemoglobin level in this age group decreased significantly from 115.7 g/L in 2007 to 113.5 g/L in 2010. Among nonpregnant mothers of children enrolled in the survey, anemia prevalence was 13.2% in 2007 (95% CI, 10.2% to 16.9%; n = 394) and 14.2% in 2009 (95% CI, 11.4% to 17.5%; n = 507). Mean hemoglobin concentration was 134.3 g/L in 2007 and 132.5 g/L in 2009. There were no statistically significant differences in anemia prevalence or mean hemoglobin levels between 2007 and 2009.

2007

2008

2009

2010

70 60 50 40 30 20 10 0

6–11

12–23

24–35

36–47

48–59

Age group (mo)

FIG. 1. Prevalence of anemia by age group and survey year

269

Micronutrient intervention and anemia in refugee children

TABLE 4. Prevalence of malnutrition and mean anthropometry scores, by time of surveya Time of survey Variable Malnutrition prevalence—% (95% CI) Acute malnutrition Moderate Severe Total

Jan 2007 (n = 497)

Oct 2008 (n = 502)

3.6 (2.2–5.8)A 0.6 (0.2–1.8)A 4.2 (2.8–6.4)A

8.2 (6.0–11.0)B 1.0 (0.4–2.3)A 9.2 (7.0–12.1)B

May 2009 (n = 568)

5.8 (4.1–8.1)A,B 1.4 (0.7–2.8)A 7.2 (5.4–9.6)B

May 2010 (n = 569)

7.7 (5.7–10.3)B 0.4 (0.1–1.3)A 8.1 (6.1–10.16)B

Stunting Moderate Severe Total

31.2 (27.2–35.5)A 8.0 (5.9–10.9)A 39.2 (34.9–43.7)A

27.9 (24.1–32.1)A 4.6 (3.0–6.9)B,C 32.5 (28.4–36.8)B

22.3 (18.9–26.0)B 5.8 (4.1–8.2)A,B 28.1 (24.5–32.0)B,C

20.2 (17.0–23.8)B 3.2 (1.9–5.1)C 23.4 (20.0–27.1)C

Underweight Moderate Severe Total

15.9 (12.9–19.5)A 5.0 (3.3–7.4)A 20.9 (17.5–24.8)A

24.1 (20.5–28.1)B 4.0 (2.5–6.2)A 28.1 (24.2–32.3)B

19.7 (16.6–80.8)A 2.8 (1.7–4.6)A,B 22.5 (19.2–26.2)A

19.2 (16.1–22.7)A 1.8 (0.9–3.3)B 20.9 (17.7–24.5)A

−0.49 ± 0.92A −1.63 ± 1.07B −1.25 ± 1.00A,B

−0.85 ± 0.93C −1.56 ± 0.91B −1.46 ± 0.93C

−0.76 ± 0.91B,C −1.47 ± 1.02B −1.35 ± 0.90B,C

−0.65 ± 0.91B −1.33 ± 0.9A −1.20 ± 0.88A

Mean ± SD z-score Weight-for-height Height-for-age Weight-for-age

a. Values in the same row not sharing the same capital superscript letter are significantly different (p < .05).

of 60%. The mean height-for-age z-score increased over the same period from –1.63 to –1.33 (table 4). Morbidity, healthcare utilization, and coverage of vitamin A and deworming programs

The reported 2-week cumulative incidence of diarrhea decreased from 30% in 2007 to 17% to 18% in 2008 and 2009 and 13% in 2010 (table 5). The 2-week cumulative incidence of acute respiratory infection increased from 28% in 2007 and 2008 to 36.3% in 2009 and then decreased to 20.4% in 2010. The percentages of children brought to the health clinic for diarrhea or acute respiratory infection remained consistently high: 75% to 80% for diarrhea and 91% to 96% for acute respiratory infection. Coverage of the vitamin A supplementation program remained consistently high at 96% to 99%, and coverage of the deworming program increased from 89% in 2007 to 95% to 97% in 2009 and 2010 (table 5). Vita-Mix-It program coverage and acceptance

The coverage, reported use, and acceptance of the Vita-Mix-It supplementation program were consistently high. The reported percentage of children who received the most recent distribution of Vita-Mix-It increased from 91% in 2008 to 97% to 98% in 2009 and 2010. Over 90% of children in each of the surveys from 2008 to 2010 were reported to currently consume

Vita-Mix-It (table 6). In the 2010 survey, 98.4% of the children had the Vita-Mix-It distribution card available. The percentage of children for whom the mother or caregiver reported any perceived negative health effects (diarrhea, vomiting, black stool, constipation, etc.) attributed to Vita-Mix-It consumption decreased progressively from 11.6% in 2008 to 5.6% in 2009 and 2.9% in 2010. Forty to fifty percent of the mothers or caregivers reported changes to food after mixing it with Vita-Mix-It. The majority of these were changes in color, when rice was reported to become slightly yellowish when mixed with Vita-Mix-It. The data from 2010 indicated positive attitudes toward Vita-Mix-It, with 80% to 85% of mothers or caregivers reporting perceived positive changes in the child’s health, energy level, and appetite after the child started receiving VitaMix-It (table 6). Breastfeeding and complementary feeding

Close to 100% of children aged 6 to 23 months were breastfed. The prevalence of breastfeeding gradually decreased after the age of 2 years. Overall, approximately 35% of children aged 24 to 59 months were breastfed. Rice and dal (lentils) were the main staple foods; Unilito consumption was also common (table 7). Reported consumption of meat and fruits increased during the study period. In 2010, 23.6% of younger (6 to 23 months) and 37.2% of older (24 to 59 months)

270

O. Bilukha et al.

TABLE 5. Morbidity, healthcare utilization, and program coverage, by time of survey Time of survey Indicator—% (95% CI)

Jan 2007 (n = 497) wka

Diarrhea in previous 2 Taken to health centerb Acute respiratory infection in previous 2 wka Taken to health centerc Received vitamin A in previous 6 mo Received albendazole in previous 6 mod a. b. c. d.

(26.0–34.3)A

Oct 2008 (n = 502)

May 2009 (n = 568)

(13.8–20.6)B,C

May 2010 (n = 569)

(15.0–21.5)B

30.0 74.5 (66.7–81.3) 28.8 (24.9–33.0)A

16.9 76.2 (65.7–84.8) 28.5 (24.6–32.7)A

18.0 80.4 (71.4–87.6) 36.3 (32.4–40.5)B

13.2 (10.6–16.3)C 80.0 (69.2–88.4) 20.4 (17.2–24.0)C

91.6 (85.8–95.6) 96.6 (94.5–97.9)

95.8 (91.0–98.4) 96.4 (94.3–97.8)

95.6 (91.9–98.0) 98.9 (97.6–99.6)

95.7 (90.2–98.6) 97.7 (96.0–98.7)

89.2 (85.9–91.9)

91.7 (88.6–94.0)

97.6 (95.7–98.7)

95.1 (92.7–96.7)

Values in the same row not sharing the same capital superscript letter are significantly different (p < .05). Percentage of those children who were reported to have had diarrhea within the previous 2 weeks. Percentage of those children who were reported to have had acute respiratory infection within the previous 2 weeks. Children < 12 mo excluded.

TABLE 6. Vita-Mix-It program indicators, by time of survey Time of survey Oct 2008 (n = 502)

Indicator—%

May 2009 (n = 568)

May 2010 (n = 569)

Distribution card availablea





98.4

Received Vita-Mix-It last distribution

91.0

98.1

97.2

Currently giving to child

90.6

97.7

91.7

Reported changes to food when mixed with Vita-Mix-Itb

44.6

39.3

56.9

Color change

41.5

37.3

52.1

Reported negative health

effectsb

11.6

5.6

2.9

Nausea and/or vomiting

4.0

1.3

0.9

Diarrhea

5.3

2.5

0.2

Black stool

0.0

1.4

0.2

Increase noticed after child received Vita-Mix-Ita Appetite Energy level Health

— — —

— — —

78.5 84.1 86.1

a. Data were only available for May 2010 b. Percentage of those who were currently giving Vita-Mix-It to the child.

children were reported to have consumed meat or fish in the preceding 24 hours. Over 50% of children consumed milk, and over 70% consumed fruits. Reported tea consumption was high in 2007 and 2008 (30% to 33% in younger children and 62% to 69% in older children) but decreased dramatically starting in 2009 and was only 1.5% in younger children and 8.3% in older children in the 2010 survey (table 7).

Discussion Strengths and limitations of the study

This study presents the first evaluation of micronutrient

powder effectiveness in a large-scale distribution program with periodic follow-up over a 26-month period, rather than in a shorter (generally, 2 to 6 months long) efficacy trial. This study evaluated program characteristics and impact in large population-representative cross-sectional samples rather than in smaller cohorts of children followed up over time. In contrast to previous studies that were limited to the age group from 6 to 24 months and used micronutrient powders containing six or fewer micronutrients, this supplementation program included a wider age range of children (6 to 59 months) and used a micronutrient powder formulation with 16 micronutrients. The results of this study should be interpreted with caution, given several important limitations. First,

271

Micronutrient intervention and anemia in refugee children

TABLE 7. Prevalence of current breastfeeding and foods consumed in the previous 24 hours, by time of survey Year of survey

Indicator Currently breastfeeding—% of children 6–23 mo 24–59 mo

May 2010 May 2009 Oct 2008 Jan 2007 (total n = 569; (total n = 568; (total n = 502; (total n = 497; 6–23 mo, n = 169; 6–23 mo, n = 160; 6–23 mo, n = 206; 6–23 mo, n = 195; 24–59 mo, 24–59 mo, 24–59 mo, 24–59 mo, n = 374) n = 362) n = 342) n = 328)

96.4 35.7

92.5 36.0

98.1 34.8

96.4 32.6

Rice or rice porridge 6–23 mo 24–59 mo

89.3 99.7

96.3 99.4

97.6 99.2

97.9 99.7

Dal 6–23 mo 24–59 mo

66.9 81.4

85.0 87.7

90.3 90.1

92.8 91.4

Unilito 6–23 mo 24–59 mo

59.8 78.0

55.6 65.5

66.5 74.9

54.9 61.8

Vegetables 6–23 mo 24–59 mo

79.3 95.4

83.8 95.0

88.8 96.1

87.2 96.8

Fruits 6–23 mo 24–59 mo

47.3 49.1

60.0 56.7

61.2 58.6

77.9 72.5

Meat or fish 6–23 mo 24–59 mo

11.2 10.4

15.0 27.2

19.4 27.6

23.6 37.2

Milk 6–23 mo 24–59 mo

45.6 41.8

55.6 49.4

56.3 56.1

57.4 56.7

Tea 6–23 mo 24–59 mo

30.8 62.2

33.1 69.0

3.9 16.0

1.5 8.3

Foods consumed in previous 24 h—% of children

because the Vita-Mix-It distribution was implemented simultaneously in all camps, the study followed a pre– post design with no age-appropriate control group. In order to monitor changes in anemia in population groups not covered by the intervention, we assessed hemoglobin status in nonpregnant women of reproductive age in 2007 and 2009. This, however, did not provide a fully adequate control group to assess the effects of the Vita-Mix-It distribution program versus other factors that may have influenced hemoglobin status in children during the study period. Therefore, it is not possible to attribute biological effects (or lack thereof) observed in this study to the intervention program alone. Second, for logistical and planning reasons,

there was no survey conducted immediately before the program started in March of 2008. We, therefore, used the most recent preintervention population-representative data collected in January and February 2007, 13 months before the start of the Vita-Mix-It program. The anemia, morbidity, and anthropometry status of children in March 2008 may have differed from those measured in 2007, and therefore the 2007 survey may not have provided the best baseline data for assessing the effects of Vita-Mix-It supplementation. Third, because of logistical and monetary constraints, we did not measure biochemical indicators of iron status (e.g., ferritin, transferrin receptor) or of the status of any other micronutrients. Using hemoglobin as a sole

272

biological indicator does not permit assessment of the changes in status of important micronutrients contributing to anemia, and thus limits the interpretability of results. Fourth, because of the logistical and monetary constraints, we were not able to measure biochemical markers of inflammation, which limits the interpretability of our results. Fifth, information reported by mothers or caregivers is subject to reporting bias. Finally, the surveys were conducted during different seasons, and therefore differences in the results may be in part due to seasonality if indicators of interest change following seasonal patterns. Only the 2009 and 2010 surveys were done at the same time of the year (early May), and therefore comparisons between these two surveys are less likely to be influenced by seasonality. Furthermore, refugees receive the same food rations throughout the year, and therefore the impact of seasonality on indicators related to or dependent on food intake is not likely to be great. Effects on hemoglobin status

Overall, our study did not find a significant reduction in the prevalence of anemia or an increase in hemoglobin levels among children aged 6 to 59 months after 26 months of Vita-Mix-It program delivery. A modest decrease in anemia prevalence observed in 2009 did not hold in 2010. The level of moderate anemia decreased from 18.9% in 2007 to 14.4% in 2010 (a relative decrease of about 24%, p < .05), and the level of severe anemia remained negligible throughout the study period. The reasons for this limited impact of Vita-Mix-It on hemoglobin levels need to be carefully considered, as the lessons learned may be critical for future program planning in this and other settings. Since anemia prevalence, iron requirements, and access to a variety of foods differ substantially between younger and older children, we will discuss the two age groups, 6 to 23 months and 24 to 59 months, separately. In order to better understand the lack of effect in older children in our study, it is important to consider the results of the study by Zlotkin and colleagues that demonstrated that 6 months of daily supplementation with micronutrient powder had no effect on hemoglobin levels in mildly anemic and nonanemic children (all with hemoglobin > 100 g/L, mean hemoglobin 112 to 113 g/L) aged 8 to 20 months at recruitment [17]. The authors proposed several explanations for this lack of effect. First, iron requirements are the highest between 6 and 12 months of age, whereas the children in the study were on average 15 to 16 months of age at the start and thus required less iron to maintain hemoglobin levels. Second, the children were mildly anemic or nonanemic (all with hemoglobin > 100 g/L) and therefore had iron absorption mechanisms down-regulated compared with moderately or severely anemic subjects. Third, the children, since they were mostly

O. Bilukha et al.

older than 12 months, had access to a wider variety of iron-rich foods than did children less than 1 year of age. When we consider refugee children aged 24 to 59 months in our study, all of the above arguments are even more relevant to them. First, these children were older (24 to 59 months) and thus had iron requirements the same as or lower than those of the children 8 to 20 months of age in the study of Zlotkin and colleagues [17]. Second, these children were mostly nonanemic or mildly anemic at the beginning of the study (the proportion of children with hemoglobin < 100 g/L was only 11%, and mean hemoglobin was 115.7 g/L in this group at baseline in 2007), and therefore iron absorption mechanisms in children were likely to be down-regulated and absorption of iron from Vita-MixIt supplements was likely to be minimal. Third, the children were older and therefore had better access to a variety of iron-rich foods (as shown in table 7). Last, the children in our study received an eight-times-lower dose of iron in their micronutrient powder supplements than did the children in the study by Zlotkin and colleagues (10 mg of iron in Vita-Mix-It once in 2 days versus 40 mg of iron daily). Therefore, the absence of effect of Vita-Mix-It on hemoglobin levels in this age group in our study is not entirely unexpected. Since to date we have found no published studies on the efficacy of micronutrient powder supplementation on anemia in children aged 24 to 59 months, caution should be exercised when planning micronutrient powder supplementation programs that include this age group. Little benefit in terms of hemoglobin increase may be expected in populations in which the prevalence of children with hemoglobin < 100 g/L is low. The limited effect of Vita-Mix-It on hemoglobin levels in the younger (6 to 23 months) age group in our study is more puzzling. The proportion of children with hemoglobin < 100 g/L in this age group was 33.7% at baseline and decreased to 27.7% in 2010—a relative decrease of about 18%, which did not reach statistical significance. It is difficult to interpret this finding in the absence of data on iron status, vitamin A status, and potentially the status of other relevant micronutrients. The lack of response may be because most children in this population have anemia that is due to nonnutritional causes, or because of inadequate intake or absorption of iron and/or other micronutrients from the supplements. Although almost all refugee children receive deworming pills every 6 months, the prevalence of hookworm infection in this population is unknown. Since this was a pre–post study without a control group, it is important to consider factors other than Vita-Mix-It supplementation that may have affected hemoglobin levels during the study period. Especially important are those factors that could have had a negative impact on hemoglobin levels and thus prevented Vita-Mix-It supplementation from achieving significant improvement in hemoglobin status. From

273

Micronutrient intervention and anemia in refugee children

the food consumption patterns (table 7), it appears that access of children to fruits, vegetables, milk, and meat either remained stable or improved during the study period, and breastfeeding practices did not change substantially. The incidence of malaria in the camps was very low, and therefore it is unlikely that malaria had a discernible effect on hemoglobin levels in children during the years from 2007 to 2010. The 2-week cumulative incidence rates of acute respiratory infection and diarrhea decreased from 2007 to 2010 and were thus unlikely to have had a negative impact on anemia. Following the results of the 2008 survey that showed no effect of Vita-Mix-It on anemia prevalence, the content of vitamin C in Vita-Mix-It was doubled from 30 mg to 60 mg in order to enhance iron absorption. At the same time, in early 2009, educational messages to reduce tea consumption in children were introduced as part of mothers’ education at enrollment into the Vita-Mix-It program. These messages appear to have had a substantial effect in reducing reported tea consumption among children, which is likely to have a positive effect on iron absorption from foods and supplements. Finally, anemia prevalence in nonpregnant mothers in 2007 and 2009 remained stable at 13% to 14%, suggesting that there were no major changes in anemia status in adult refugees. In summary, we could not identify any major factors that could have had a substantial negative impact on hemoglobin levels in children during the study period. Therefore, it is important to investigate the extent to which non-nutritional causes of anemia affected this population and to assess other indicators of micronutrient status, in particular vitamin A and iron status. Effects on growth

Another important finding of this study was a dramatic decrease in stunting. The prevalence of stunting decreased progressively during the study period from 39.2% in 2007 to 23.4% in 2010, a relative decrease of about 40%. The mean height-for-age z-score improved over the same period from –1.63 to –1.33. However, since this was a pre–post evaluation without a control group, it is not clear how much of this improvement could be attributed to the Vita-Mix-It supplementation. Other factors, such as decrease in morbidity, good health program coverage and high healthcare utilization, improved access to nutritious foods, and improved complementary feeding practices, could have all contributed to improved linear growth. A large magnitude of decrease in stunting was unexpected to some extent, since it was not among the a priori expected outcomes of the micronutrient powder intervention but was rather included in the surveys as part of routine nutrition monitoring. Further, most previous micronutrient supplementation studies did not find a substantial effect on growth. In the recent

systematic review [27], the authors identified six efficacy trials that evaluated the effects of micronutrient fortification of complementary foods on child growth. Two of these studies used micronutrient powder [14, 28] consisting of six micronutrients (iron, zinc, folate, and vitamins A, C, and D); the other studies used fortified cereals, milk, or micronutrient tablets. Only one of the six efficacy trials, a milk-fortification study in India [29], found a significant improvement in heightfor-age at the end of the 12-month supplementation period among children 12 to 36 months of age. None of the other five studies, including two efficacy trials that used micronutrient powder, showed an effect on linear growth. Similarly, a large multicenter study that tested the effects of daily and weekly micronutrient supplementation in the form of chewable tablets (foodlets) in infants aged 6 to 11 months in Indonesia, Peru, South Africa, and Vietnam found no significant effect on height-for-age z-scores [30]. Children in this study received tablets that included 15 micronutrients (similar in composition to Vita-Mix-It) for a period of 6 months. The meta-analysis of studies comparing the effects of supplementation with three or more micronutrients with the effects of supplementation with one or two micronutrients or with placebo found a small, albeit significant, positive effect on linear growth [31]. Although the decrease in stunting found in our study cannot be clearly attributed to Vita-Mix-It supplementation, and although most previous studies have not shown similar effects on linear growth, the findings of our study should not be dismissed entirely. The length of supplementation, stunting levels at baseline, age of the children enrolled in the program, and composition of supplements, among other factors, may account for differences in effects. More studies are needed to confirm or disprove the possible beneficial effects of multimicronutrient powders on linear growth. The reasons for the increases in the prevalence of wasting and underweight seen in 2008 and the subsequent decreases in these indicators in 2009 and 2010 are not well understood and could not be clearly attributed to Vita-Mix-It supplementation. Effects on morbidity

The 2-week cumulative incidence of diarrhea decreased from 30% in 2007 to 17% to 18% in 2008 and 2009 and to 13% in 2010. The 2-week cumulative incidence of acute respiratory infection increased from 28% to 29% in 2007 and 2008 to 36% in 2009 and then decreased to 20% in 2010. A large decrease in acute respiratory infection from 2009 to 2010 is especially interesting, since these surveys were conducted during the same season, and thus seasonality patterns were not likely to affect the results. It is possible that decreases in diarrhea and acute respiratory infection may be due in part to Vita-Mix-It supplementation. Of five efficacy

274

trials that evaluated the effects of complementary food fortification on morbidity included in a recent systematic review [27], two [32, 33] reported significant decreases in the incidence of diarrhea and fever in the group receiving micronutrient supplements compared with controls. Vita-Mix-It program delivery and acceptance

Data on Vita-Mix-It distribution program coverage and Vita-Mix-It use collected from 2008 to 2010 consistently show very high coverage and compliance. Coverage data collected in the surveys matched the distribution data from nutrition units in the camps, which also showed over 95% coverage. In 2010, when we were able to document coverage using distribution cards rather than self-report, 99% of children had cards available and over 97% of children received the last distribution according to the cards. Acceptance of Vita-Mix-It supplements among mothers and children also appeared to be consistently high, with 80% to 85% of mothers and caregivers reporting that Vita-Mix-It had positive effects on their children’s health, energy level, and appetite. The percentage of those who reported perceived negative health effects on their child attributed to Vita-Mix-It decreased from 12% in 2008 to only 3% in 2010. Close to half of the participants reported that rice became slightly yellowish in color when mixed with Vita-Mix-It, but this change did not to seem to have any substantial effect on acceptance or compliance. In general, the Vita-Mix-It distribution program in Bhutanese refugee camps may serve as a model for a well-planned, well-implemented, and well-monitored program. Moreover, the necessary adjustments were made according to data collected through monitoring and evaluation. For example, the vitamin C content in Vita-Mix-It was doubled, and messages on appropriate complementary feeding practices and negative effects of tea in children were added to the mothers’ education sessions based on the results of the survey conducted in 2008. Vita-Mix-It distribution cards were introduced in 2009 to better track coverage. The above data on program coverage and acceptance do not suggest that lack of compliance may have been a contributing factor to the lack of effectiveness of Vita-Mix-It supplementation on hemoglobin status. However, data on consumption collected in our surveys are subject to reporting bias, and more persuasive data can be collected by household visits that would permit direct observation of use.

Conclusions Our study presents the first evidence from a multiyear, longitudinal follow-up of a large-scale micronutrient powder distribution program targeting children aged

O. Bilukha et al.

6 to 59 months in refugee settings. The program was well implemented and achieved high coverage, and the caregivers reported positive attitudes and high use of supplements. One of the most important findings was the limited effect on hemoglobin levels and anemia prevalence. Although a low impact in older children (aged 24 to 59 months) could have been anticipated to a certain degree, given the low baseline prevalence of moderate anemia in this age group, the limited effect in younger children (6 to 23 months) is more puzzling and requires further investigation and discussion. Including additional biochemical indicators to better assess the status of iron and other micronutrients, and investigating the extent to which anemia in this population is due to non-nutritional causes, would probably provide additional valuable information to address this question. Another important finding of this study was a large reduction in the prevalence of stunting. Although the design of this evaluation cannot confirm that this reduction was due to Vita-Mix-It supplementation, the results are clearly noteworthy. More data are needed, especially from multiyear, large-scale supplementation programs that include multiple micronutrients and older age groups, in order to refute or confirm a causal effect of micronutrient powder supplementation on stunting. The large decrease in diarrhea morbidity is also an important finding. Both the decreases in stunting and in diarrhea suggest that other micronutrients (zinc in particular) included in this micronutrient powder formulation may have had a substantial positive effect on child health and development. Lessons learned from this and other studies included in this issue should provide invaluable information both for formulating the agenda for future research on micronutrient powder effectiveness in large-scale program settings and for program planning, implementation, and monitoring.

Disclaimer The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention or the Office of the United Nations High Commissioner for Refugees.

Acknowledgments We are grateful to the field staff of the Association of Medical Doctors of Asia, the Office of the United Nations High Commissioner for Refugees, the World Food Programme, and the Bhutanese refugee health and nutrition workers who participated in this study.

275

Micronutrient intervention and anemia in refugee children

References 1. McLean E, Egli I, Cogswell M, de Benoist B, Wojdyla D. Worldwide prevalence of anemia in preschool aged children, pregnant women and non-pregnant women of reproductive age. In: Kraemer K, Zimmerman MB, eds. Nutritional anemia. Basel, Switzerland: Sight and Life, 2007:1–12. 2. Black RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzafi M, Mathers C, Rivera J, for the Maternal and Child Undernutrition Study Group. Maternal and child undernutrition: Global and regional exposures and health consequences. Lancet 2008;371:243–60. 3. Scrimshaw NS. Functional significance of iron deficiency: An overview. In: Enwonwu CO, ed. Annual Nutrition Workshop Series, Vol III. Functional significance of iron deficiency. Nashville, Tenn, USA: Meharry Medical College, 1990:1–13. 4. Grantham-McGregor S, Ani C. A review of studies on the effect of iron deficiency on cognitive development in children. J Nutr 2001;131:649S–68S. 5. Woodruff BA, Blanck HM, Slutsker L, Cookson ST, Larson MK, Duffield A, Bhatia R. Anaemia, iron status and vitamin A deficiency among adolescent refugees in Kenya and Nepal. Public Health Nutr 2006;9:26–34. 6. Kemmer TM, Bovill ME, Kongsomboon W, Hansch SJ, Geisler KL, Cheney C, Shell-Duncan BK, Drewnowski A. Iron deficiency is unacceptably high in refugee children from Burma. J Nutr 2003;133:4143–9. 7. Seal AJ, Creeke PI, Mirghani Z, Abdalla F, McBurney RP, Pratt LS, Brookes D, Ruth LJ, Marchand E. Iron and vitamin A deficiency in long-term African refugees. J Nutr 2005;135:808–13. 8. Zlotkin SH, Schauer C, Christofides A, Sharieff W, Tondeur MC, Hyder SMZ. Micronutrient sprinkles to control childhood anemia. PLoS Med 2005;2:e1. 9. Zlotkin S, Arthur P, Schauer C, Antwi KY, Yeung G, Piekarz A. Home-fortification with iron and zinc sprinkles or iron sprinkles alone successfully treats anemia in infants and young children. J Nutr 2003;133:1075–80. 10. Zlotkin S, Arthur P, Antwi KY, Yeung G. Treatment of anemia with microencapsulated ferrous fumarate plus ascorbic acid supplied as sprinkles to complementary (weaning) foods. Am J Clin Nutr 2001;74:791–5. 11. Hirve S, Bhave S, Bavdekar A, Naik S, Pandit A, Schauer C, Christofides A, Hyder Z, Zlotkin S. Low dose ‘Sprinkles’—an innovative approach to treat iron deficiency anemia in infants and young children. Indian Pediatr 2007;44:91–100. 12. Christofides A, Asante KP, Schauer C, Sharieff W, Owsu-Agyei S, Zlotkin S. Multi-micronutrient Sprinkles including a low dose of iron provided as microencapsulated ferrous fumarate improves hematologic indices in anaemic children: A randomized clinical trial. Matern Child Nutr 2006;2:169–80. 13. Adu-Afarwuah S, Lartey A, Brown KH, Zlotkin S, Briend A, Dewey KG. Home fortification of complementary foods with micronutrient supplements is well accepted and has positive effects on infant iron status in Ghana. Am J Clin Nutr 2008;87:929–38. 14. Giovannini M, Sala D, Usuelli M, Livio L, Francescato G, Braga M, Radaelli G, Riva E. Double-blind,

15.

16.

17.

18.

19. 20.

21. 22. 23.

24.

25.

26.

27.

placebo-controlled trial comparing effects of supplementation with two different combinations of micronutrients delivered as sprinkles on growth, anemia, and iron deficiency in Cambodian infants. J Pediatr Gastroenterol Nutr 2006;42:306–12. Hyder SMZ, Haseen F, Rahman M, Tondeur MC, Zlotkin SH. Effect of daily versus once-weekly home fortification with micronutrient Sprinkles on hemoglobin and iron status among young children in rural Bangladesh. Food Nutr Bull 2007;28:156–64. Ip H, Hyder SMZ, Haseen F, Rahman M, Zlotkin SH. Improved adherence and anaemia cure rates with flexible administration of micronutrient Sprinkles: A new public health approach to anaemia control. Eur J Clin Nutr 2009;63:165–72. Zlotkin S, Antwi KY, Schauer C, Yeung G. Use of microencapsulated iron(II) fumarate sprinkles to prevent recurrence of anemia in infants and young children at high risk. Bull World Health Organ 2003;81:108–15. Menon P, Ruel MT, Loechl CU, Arimond M, Habicht JP, Pelto G, Michaud L. Micronutrient Sprinkles reduce anemia among 9- to 24-mo-old children when delivered through an integrated health and nutrition program in rural Haiti. J Nutr 2007;137:1023–30. Centers for Disease Control and Prevention. Malnutrition and micronutrient deficiencies among Bhutanese refugee children—Nepal, 2007. MMWR 2008;57:370–3. World Health Organization/World Food Programme/ United Nations Children’s Fund. Preventing and controlling micronutrient deficiencies in populations affected by an emergency. Multiple vitamin and mineral supplements for pregnant and lactating women, and for children aged 6 to 59 months. Geneva: WHO/WFP/ UNICEF, 2007. Food and Agriculture Organization/World Health Organization. Vitamin and mineral requirements in human nutrition, 2nd ed. Geneva: WHO, 2005. Cohen AR, Seidl-Friedman J. HemoCue system for hemoglobin measurement: Evaluation in anemic and non-anemic children. Am J Clin Pathol 1988;90:302–5. World Health Organization/United Nations High Commissioner for Refugees/International Food Policy Research Institute/World Food Programme. The management of nutrition in major emergencies. Geneva: WHO, 2000. United Nations Department of Technical Co-operation for Development and Statistical Office. How to weigh and measure children. Assessing the nutritional status of young children in household surveys. New York: United Nations, 1986. WHO child growth standards: Length/height-for-age, weight-for-age, weight-for-height and body mass indexfor-age: Methods and development. Geneva: World Health Organization, 2006. International Emergency and Refugee Health Branch, CDC. Epi Info/ENA Software. Available at: http:// www.cdc.gov/globalhealth/ierh/ResearchandSurvey/ enasoftware.htm . Dewey KG, Adu-Afarwuah S. Systematic review of the efficacy and effectiveness of complementary feeding

276 interventions in developing countries. Matern Child Nutr 2008;4:24–85. 28. Adu-Afarwuah S, Lartey A, Brown KH, Zlotkin S, Briend A, Dewey KG. Randomized comparison of 3 types of micronutrient supplements for home fortification of complementary foods in Ghana: Effects on growth and motor development. Am J Clin Nutr 2007;86:412–20. 29. Dhingra P, Menon VP, Sazawal S, Dhingra U, Marwah D, Sarkar A. Effect of fortification of milk with zinc and iron along with vitamins C, E, A and selenium on growth, iron status and development in pre-school children—a community-based double-masked randomized trial. Report from the 2nd World Congress of Pediatric Gastroenterology, Hepatology and Nutrition. Paris, 3–7 July 2004. 30. International Research on Infant Supplementation (IRIS) Study Group: Smuts CM, Lombard CJ, Spinnler Bernade AJ, Dhansay MA, Berger J, Hop LT, Lopez de

O. Bilukha et al.

Romana G, Untoro J, Karyadi E, Erhardt J, Gross R. Efficacy of a foodlet-based multiple micronutrient supplement for preventing growth faltering, anemia, and micronutrient deficiency of infants: The four country IRIS trial pooled data analysis. J Nutr 2005;135:631S–8S. 31. Allen LH, Peerson JM, Olney DK. Provision of multiple rather than two or fewer micronutrients more effectively improves growth and other outcomes in micronutrientdeficient children and adults. J Nutr 2009;139:1022–30. 32. Sazawal S, Dhingra U, Hiremath G, Kumar J, Dhingra P, Sarkar A, Menon VP, Black RE. Effects of fortified milk on morbidity in young children in north India: Community based, randomised, double masked placebo controlled trial. BMJ 2007;334:140–3. 33. Sharieff W, Bhutta Z, Schauer C, Tomlinson G, Zlotkin S. Micronutrients (including zinc) reduce diarrhoea in children: The Pakistan Sprinkles Diarrhoea Study. Arch Dis Child 2006;91:573–9.

Provision of micronutrient powder in response to the Cyclone Sidr emergency in Bangladesh: Crosssectional assessment at the end of the intervention

Jee Hyun Rah, Saskia de Pee, Siti Halati, Monira Parveen, Syeda Sajia Mehjabeen, Georg Steiger, Martin W. Bloem, and Klaus Kraemer Abstract Background. Micronutrient powder has been endorsed as an effective means to improve the micronutrient status of emergency-affected populations. Objective. To document the experience and findings of a cross-sectional assessment of the micronutrient powder program implemented as part of the emergency response to Cyclone Sidr. Methods. Micronutrient powder was distributed to 100,714 children under 5 years of age and 59,439 pregnant or lactating women severely affected by Cyclone Sidr in Bangladesh. A cross-sectional assessment, including hemoglobin and anthropometric measurements, was conducted after the completion of the micronutrient powder program among children under 5 years of age, lactating mothers, and postmenarcheal adolescent girls in the intervention area. Comparison groups for each, drawn from the control area, which had not received micronutrient powder, were assessed at the same time. Results. The prevalence of anemia among children under 5 years of age was approximately 80% in both areas. Among children in the intervention area, those who consumed at least 75% of the micronutrient powder sachets had a lower prevalence of stunting than those who consumed less than 75% of the sachets (40% vs. 52%, p < .05). Among lactating mothers in the intervention Jee Hyun Rah and Klaus Kraemer are affiliated with Sight and Life, Basel, Switzerland; Saskia de Pee is affiliated with the World Food Programme, Rome, and the Friedman School of Nutrition Science and Policy, Tufts University, Boston, Massachusetts, USA; Siti Halati is affiliated with the World Food Programme, Nepal; Monira Parveen and Syeda Sajia Mehjabeen are affiliated with the World Food Programme, Dhaka, Bangladesh; Georg Steiger is affiliated with DSM Nutritional Products, Basel, Switzerland; Martin W. Bloem is affiliated with the World Food Programme, Rome, the Friedman School of Nutrition Science and Policy, Tufts University, Boston, Massachusetts, USA, and the Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA. Please direct queries to the corresponding author: Jee H. Rah, Sight and Life, Wurmisweg 576, B 241/313, 4303, Kaiseraugst, Switzerland; e-mail: [email protected].

area, the prevalence rates of thinness and anemia were lower among those who consumed at least 75% of the sachets than among those who consumed less than 75% of the sachets (thinness, 31% vs. 46%, p < .05; anemia, 50% vs. 61%, p = .07). For adolescent girls in the intervention and control areas, none of whom had received micronutrient powder, the prevalence rates of anemia were 52% and 45%, respectively (p = .05). Conclusions. Micronutrient powder may reduce anemia among lactating mothers, when the compliance rate is high. Anemia prevalence prior to micronutrient powder distribution had not been investigated and could have been higher among children and lactating mothers in the intervention than in the control area, resulting in the negation of the potential positive impact of micronutrient powder on anemia.

Key words: Anemia, Bangladesh, Cyclone Sidr, emergency, micronutrient powder, nutritional status Introduction Super Cyclonic Storm Sidr hit 15 southern coastal districts of Bangladesh on 15 November 2007, causing nearly 10,000 deaths and large-scale evacuations. The Bangladesh Country Office of the United Nations World Food Programme (WFP) immediately responded to the disaster by providing general food rations to an estimated 2.2 million people in nine cyclone-affected districts. These food rations included rice, pulses, vitamin A– and D–fortified vegetable oil, iodized salt, and micronutrient-fortified wheat–soy blend. The distribution of multiple micronutrient powder, in addition to general food rations, was started in mid-2008 to provide additional micronutrients to specific target groups among the beneficiaries in order to reduce the prevalence and severity of micronutrient deficiencies. In emergency conditions, the risk of micronutrient deficiency increases due to pre-existing deficiencies

Food and Nutrition Bulletin, vol. 32, no. 3 © 2011, The United Nations University.

277

278

and infectious morbidities such as diarrhea among vulnerable groups. Moreover, the quality of their diet is likely to be suboptimal, lacking essential micronutrients and further worsened by inconsistent and interrupted food supplies. Therefore, it is imperative to ensure that the micronutrient needs of people affected by disasters are adequately met [1]. Micronutrient powder, in this case branded as MixMe and locally branded as Pushtika (meaning “something that is nutritious”), is a food supplement packaged in single-dose sachets. These sachets contain micronutrients in powdered form and can be added to foods prepared in the household just before consumption. In general, one dose consists of 1 g of powder, which contains one Recommended Nutrient Intake (RNI) of each of the vitamins and minerals contained [2]. In 2007, the use of micronutrient powder, particularly in emergency conditions, was endorsed by the World Health Organization (WHO), WFP, and UNICEF in a joint statement as an effective way of improving the micronutrient status of the nutritionally vulnerable sections of the population, such as children under 5 years of age and pregnant and lactating women [1]. Numerous studies have tested the efficacy of micronutrient powder, primarily among young children in controlled settings, and demonstrated that micronutrient powder is efficacious in the treatment and prevention of iron-deficiency anemia [3]. The effectiveness of micronutrient powder in large-scale program conditions has been demonstrated mostly in development settings [4], and also in an emergency setting in Aceh, Indonesia [5]. In the present program, the distribution of Pushtika aimed to reach approximately 110,000 children under 5 years of age and 55,000 pregnant and lactating women out of half the population of 2.2 million receiving general food rations in the nine cyclone-affected districts. After the completion of the micronutrient powder intervention, a cross-sectional survey, including hemoglobin and anthropometric measurements, was conducted among children 12 to 59 months of age and lactating mothers who had received micronutrient powder and adolescent girls who had not received micronutrient powder in the implementation areas. In addition, children under 5 years of age, lactating women, and adolescent girls from the 1.1 million beneficiaries of WFP who were not selected for the micronutrient powder distribution were interviewed and assessed for the purpose of comparison. The aim of the cross-sectional assessment was to examine how the provision of micronutrient powder, in response to the Cyclone Sidr emergency, affected nutritional status, as measured by anthropometry and anemia prevalence in the target beneficiaries. The focus of the present paper is to document the experience and findings of this cross-sectional assessment of the micronutrient

J.-H. Rah et al.

powder program implemented as part of the WFP emergency response.

Methods Program implementation

The micronutrient powder program was implemented from August 2008 to the end of January 2009 covering 100,714 children 6 to 59 months of age and 59,439 pregnant and lactating women in the four southern coastal districts severely affected by Cyclone Sidr, namely Barguna, Bagerhat, Patuakhali, and Pirojpur. The targeted children and women also received general food assistance. The micronutrient formulation of Pushtika was developed so that one sachet provided 1 RNI of each of 15 micronutrients for children 1 to 3 years of age (table 1) [2, 6]. Two sachets provided an amount close to 1 RNI for pregnant and lactating women. The formulation was in line with the WHO/WFP/UNICEF joint statement, with some downward adjustment made for vitamin A and iodine, taking into account fortification levels of the fortified oil, wheat–soy blend, and salt in the general food rations [1]. In consideration of the increased nutrient needs during pregnancy, micronutrient powder was provided in addition to the regular provision of iron and folic acid supplementation for pregnant women (the iron and folic acid supplement was also provided in the control area). The recommended micronutrient powder dosage was one sachet every other day for children under 5 years of age and two sachets every other day for pregnant and lactating women. Thus, the program targeted to provide the beneficiaries with approximately half the RNI of micronutrients every day through regular micronutrient powder intake. This dosage took into account the micronutrient content of the general food rations. Targeted children received one box with 100 sachets and pregnant and lactating women received two boxes with a total of 200 sachets at one time in August 2008, which was intended to be sufficient for a period of 6 or 7 months. The micronutrient powder distribution was implemented by two nongovernmental organization implementing partners: Save the Children USA and Proshika. The distribution could not begin immediately after the disaster, because supplies were not readily available, the micronutrient composition had to be decided and agreed upon by all partners, and permission needed to be obtained from the Government of Bangladesh for its formulation, distribution, and impact assessment. Prior to the distribution of micronutrient powder, a social marketing campaign was undertaken to raise the awareness of micronutrients and the use of micronutrient powder among the target beneficiaries. This was

279

Micronutrient powder use in Cyclone Sidr-affected areas

TABLE 1. Micronutrient composition of Pushtika

Nutrient

Unit

Vitamin A Vitamin D Vitamin E Thiamine Riboflavin Niacin Pyridoxine Vitamin B12 Folic acid Vitamin C Iron Zinc Copper Selenium Iodine

μg RE μg mg mg mg mg mg μg μg mg mg mg mg μg μg

Amount per 1-g sachet

% RNI for children 1–3 yra

WHO/WFP/ UNICEF recommendationb

100.0c 5.0 5.0 0.5 0.5 6.0 0.5 0.9 150.0 30.0 10.0 4.1 0.34d 17.0 30.0c

25 100 100 100 100 100 100 100 167 100 100 100 100 100 33

400.0 5.0 5.0 0.5 0.5 6.0 0.5 0.9 150.0 30.0 10.0 4.1 0.56 17.0 90.0

RE, retinol equivalent a. Recommended Nutrient Intake according to WHO/FAO [2]. b. WHO/WFP/UNICEF recommendation for children < 5 yr in nonmalaria area with no fortified food available [1]. c. Reduced because fortified food provided by WFP already contributes a considerable amount. d. Reduced to US RDA because the upper limit for intake by 1- to 3-year-olds is 1 mg.

necessary because the concept of home fortification and the micronutrient powder product were new to the population. As part of the social marketing campaign, training materials, including posters, flyers, and flip charts, were developed. These materials provided people with general information regarding health and nutrition and the benefits of micronutrient powder use. The personnel trained for the campaign were micronutrient powder distributors and other health and nutrition staff of the implementing partners; government health staff, including doctors, midwives, and nutritionists; and health volunteers from local clinics and village health posts. In addition, community-based education sessions were carried out on a regular basis. Cross-sectional assessment conducted at the end of the micronutrient powder intervention

Within the four districts affected by Cyclone Sidr that received micronutrient powder from WFP in addition to general food rations, the cross-sectional assessment at the end of the micronutrient powder program was carried out in two distinct areas, one that was provided with micronutrient powder and another that received general food rations but not the micronutrient powder, thus serving as the control. There were several unions (administrative units smaller than subdistricts) within the subdistricts that did not receive micronutrient powder and whose socioeconomic status, food security, and nutritional status were comparable to those of the

micronutrient powder–receiving areas. These unions were selected as the control areas, while other unions in the same subdistricts receiving micronutrient powder were selected as the intervention areas. The selected control unions were located relatively far from the intervention unions. The target respondents in the intervention area were lactating mothers who had received Pushtika, mothers of children 12 to 59 months of age who had received Pushtika, and postmenarcheal adolescent girls aged 13 to 17 years who had not received Pushtika. Comparison groups for each were drawn from the control areas. Although adolescent girls were not eligible to receive micronutrient powder, they were included in the present assessment based on the assumption that there would be no difference in anemia prevalence in adolescent girls between the two areas. If differences in anemia prevalence among children and lactating mothers in the intervention and control areas existed, the comparison of their findings with those from the adolescent girls would allow conclusions to be drawn about the impact of Pushtika with more certainty. The sample size was calculated using α = .05 and β = .2 to detect minimum differences in anemia prevalence of 15% for children and lactating mothers. The anemia prevalence in the control area was assumed to be 65% among children under 5 years of age and 45% among lactating mothers. After adjustment for a design effect of 1.3 and a 20% dropout rate, the required sample size was 350 per group in both areas. Because of logistic

280

and budget considerations, the same sample size of 350 per group was used for adolescent girls. Thus, the total required sample included 700 lactating mothers, 700 children 12 to 59 months of age, and 700 postmenarcheal adolescent girls 13 to 17 years of age. Sampling was conducted in two stages. In the first stage, all the villages in the selected unions were listed. A starting point in the list was chosen at random by a lottery method. Choices were made thereafter at regular intervals until a total of 35 villages had been selected from each area. The second stage of sampling was conducted by using a simple random sampling method in both intervention and control areas. For this, the survey team spun a pencil on the ground at the approximate center of the village to determine a random direction to be followed. Beginning with the first household from the center of the village in the selected direction, households located within a radius of 45 degrees toward the edge of the cluster were interviewed until 10 lactating mothers, 10 children under 5 years of age, and 10 adolescent girls had been selected from each village. The eligible respondents in each selected household were interviewed at home. The adolescent girls identified in each household were asked whether they had had their first menstruation, and the assessment was conducted only among the postmenarcheal girls. Prior to the interview, written consent was sought from each respondent. For adolescent girls and children under 5 years of age, consent was obtained from the parent or guardian. Assessment was conducted only after consent had been obtained. Anthropometric measurements, including length, height, and weight, were performed, and the hemoglobin levels of the lactating mothers, children, and adolescent girls were assessed. Anthropometric measurements were taken following standard procedures described by Gibson [7]. Height or length was measured to the nearest 0.1 cm with a wooden heightand-length board locally manufactured based on the UNICEF height board design. Weight was measured to the nearest 0.1 kg with a Tanita digital scale. Hemoglobin was measured by HemoCue HB 201+. Immunization cards or home records of date of birth, if available, were used to determine the age of the children. When these documents were unavailable, the mother’s recall was used or the local events calendar was used to elicit the age. Various probes, such as age at menarche, age at school enrollment, and tools such as the local events calendar, were used to determine the age of adolescents. Subsequently, a survey was conducted among lactating mothers, mothers of children 12 to 59 months of age, and mothers of adolescent girls inquiring about their demographic profile and household socioeconomic status. Information on dietary intake, morbidity, and the use of micronutrient powder by lactating

J.-H. Rah et al.

mothers, children 12 to 59 months of age, and adolescent girls was also collected. Information on the intake of various food items in the previous 7 days and morbidity in the previous 2 weeks was obtained by the recall method. Compliance was assessed by asking the target beneficiaries to recall the total number of sachets consumed. Although mothers of adolescent girls were the primary respondents to the survey, information on dietary intake, morbidity, and possible consumption of micronutrient supplements was collected directly from the adolescents. Mothers, children, and adolescent girls who were found to be severely anemic (hemoglobin < 8 g/dL) were referred to the nearest health clinic for treatment. All interviews and measurements were carried out by four teams of trained data collectors, each composed of one supervisor, four interviewers who conducted the survey and obtained anthropometric measurements, and one data collector who measured hemoglobin levels. Most of the data collectors were college graduates with prior experience in collecting survey and anthropometric data at the household level. Those who collected data on hemoglobin levels had training in measuring hemoglobin in children and women. All data collectors received 6 days of training in conducting the survey, obtaining anthropometric measurements, and assessing hemoglobin levels. Before the final selection, the data collectors had to pass a “mock test” covering the whole interview and measurement process. During the cross-sectional assessment, each team visited one village per day, and each collector was allowed to do a maximum of five or six interviews a day in order to ensure the quality of the data. The cross-sectional assessment was carried out by the Nielsen Company in Bangladesh. The protocol was approved by the Bangladesh Medical Research Council, Dhaka. The cross-sectional assessment was carried out from February to March 2009. Statistical analyses

Body mass index (BMI) was defined as the weight in kilograms divided by the square of the height in meters. Stunting and wasting in children were defined as height-for-age and weight-for-height z-scores < –2, respectively, using the new WHO growth standards in AnthroPlus 2009 software. Anemia was defined as a hemoglobin level below 11 g/dL for children under 5 years of age and 12 g/dL for lactating mothers and adolescent girls. Continuous variables were compared between intervention and control areas by t-tests, and proportions between the two areas were compared by chi-square tests. Lactating women and mothers of all groups were combined because there were no differences in demographic, socioeconomic, or other characteristics between survey respondents in the intervention

281

Micronutrient powder use in Cyclone Sidr-affected areas

and the control area for all three groups of women. The Kruskal-Wallis test was used to compare the means of continuous variables that were not normally TABLE 2. Demographic, socioeconomic, and other characteristics of survey respondents in the intervention and control areas Intervention (n = 746)

Control (n = 696)

30.7 ± 9.3 84.3a

30.9 ± 9.1 77.4

Highest grade completed (%)b 1–5 6–9 ≥ 10

63.9a 30.4 5.7a

56.6 34.0 9.5

Occupation (%) Housewife

85.1

88.9

Marital status Currently married Divorced or separated Widowed

97.1 1.2 1.6

96.9 1.0 2.0

Religion (%) Muslim Hindu Other

91.6 7.6a 0.8

89.4 10.6 0.0

Monthly income (%)d < 3,000 Tk 3,000–5,000 Tk 5,001–8,000 Tk 8,001–10,000 Tk > 10,000 Tk

10.1 52.0 26.5 5.2 6.2

10.6 53.7 27.3 4.2 4.2

Ownership (%) Radio Television Electricity

12.9a 12.9 21.2a

8.2 13.4 12.5

84.9a

90.8

Main income earner (%) Respondent’s husband

88.3

87.1

Occupation of main income earner (%) Skilled laborer Unskilled laborer Small business Farmer Other

10.7a 44.1 19.7a 8.5a 17.0

7.0 43.0 14.4 16.1 19.5

Affected by cyclone (%)

98.9

99.4

Received food ration (%)

86.9

90.1

Characteristic Mean ± SD age (yr) Ever been enrolled in school (%)

(%)c

Tube well main drinking water source (%)

a. b. c. d.

Significantly different from control group (p < .05, chi-square test). Among 1,168 women who had ever been enrolled in school. Information on marital status was missing for 25 women. 1 Tk = US$0.01.

distributed. Multiple logistic regression analyses were conducted with anemia prevalence as the dependent variable and the potential confounders as the independent variables. In order to identify potential confounding factors, characteristics such as dietary intake and economic variables that are known to be associated with anemia prevalence were identified at the outset. Bivariate analyses were then conducted for all potential risk factors by chi-square tests, and only factors that were significantly associated with anemia prevalence (p < .05) and differed between the intervention and control areas were included in the final models. All analyses were performed with STATA, version 8.0.

Results A total of 1,442 women, including lactating mothers (n = 358, intervention area; n = 361, control area), mothers of children 12 to 59 months of age (n = 429; n = 428), and mothers of adolescent girls 13 to 17 years of age (n = 346, n = 351), were interviewed during the assessment. Some women contributed to the assessment both as a lactating mother and as a mother of a child under 5 years of age and/or an adolescent girl. Most of the women were married, Muslim, full-time housewives (table 2). Approximately three-quarters of the households in both areas had a monthly income between 3,000 and 8,000 Taka (approximately US$44 to US$117). In more than half of the households in both areas, the main income earner worked as a skilled or unskilled laborer. Almost all of the mothers (99%) reported that their households were affected by Cyclone Sidr, and most received general food rations. There were no differences between the intervention and control areas in the proportions of households affected by Cyclone Sidr (99% vs. 99%), receiving food rations (87% vs. 90%), or receiving other aid (44% vs. 45%). Children 12 to 59 months of age

The mean (± SD) age of the children was 24.5 ± 11.9 months and 23.3 ± 11.4 months in the intervention and control areas, respectively. Approximately half were male in both areas. The consumption of nutritious foods such as milk, meat, and fruits was rare: less than one-third of the children had consumed any of these food items at least once in the preceding week, indicating their low dietary diversity. The proportion of children who had had diarrhea in the previous 2 weeks was low at approximately 10%, but nearly half the children in both areas had had fever and nasal discharge in the previous 2 weeks. There was no difference between the intervention and control areas in the prevalence of morbidity among the children. Approximately 42% of the children were stunted

282

J.-H. Rah et al.

and 12% were wasted; the prevalence rates did not differ between the two areas. However, among the children in the intervention area, those who consumed at least 75% of the micronutrient powder sachets had a significantly lower prevalence of stunting than those who consumed less than 75% of the sachets (p < .05) (table 3). On the other hand, the prevalence of wasting was higher among children who consumed at least 75% of the sachets than among those who consumed less, although the difference was not significant. Anemia was highly prevalent, affecting approximately 80% of the children in both areas, but there was no difference between the intervention and the control areas in the prevalence of anemia (79% and 82%, respectively). In a multivariate analysis, the prevalence of anemia did not differ between children in the intervention and the control areas after adjustment for the child’s age and intake of milk, meat, eggs, and breastmilk (data not shown). Nearly all children in the intervention area were reported to have consumed Pushtika, whereas none in the control area had consumed it. The mean (± SD) total number of sachets consumed by children was 90 ± 34, resulting in a mean adherence rate of 90%, and nearly 80% consumed more than 75% of the 100 sachets distributed in August 2008. Lactating mothers

The mean (± SD) age of the lactating mothers was 25.9 ± 5.2 years in the intervention area and 26.4 ± 5.8 years in the control area. Approximately 20% of lactating mothers reported having had symptoms such as fever, nasal discharge, and cough in the previous 2 weeks, but the proportion of mothers reporting symptoms did not differ between the two areas. The mean weight, TABLE 3. Prevalence of undernutrition among children in the intervention area according to rate of compliance with micronutrient powder usea Percentage and number of Pushtika sachets consumedb

Variable Mean ± SD age (mo) Stunting (%) Underweight (%) Wasting (%) Anemia (%) Mean ± SD hemoglobin (g/dL)

1%–74% (< 75 sachets) (n = 84)

≥ 75% (≥ 75 sachets) (n = 303)

24.3 ± 11.3 52.4 29.8 6.0 76.1 9.88 ± 1.43

24.9 ± 11.8 40.3c 30.4 12.9 80.2 9.87 ± 1.36

a. Estimated according to the 2006 WHO growth reference [10]. b. Information on the total number of sachets consumed was missing for 30 of the 417 children who had ever consumed Pushtika. c. Significantly less than the value for children who consumed fewer than 75 sachets (p < .05, chi-square test).

height, and BMI of lactating mothers did not differ between the two areas (44.9 ± 6.9 vs. 44.4 ± 6.8 kg, 151.2 ± 5.5 vs. 151.0 ± 5.5 cm, and 19.6 ± 2.7 vs. 19.5 ± 2.5 kg/m2, respectively, in the intervention and control areas). Approximately 40% of the mothers in both areas were thin (BMI < 18.5 kg/m2), reflecting chronic energy deficiency. Among the lactating mothers in the intervention area, the proportion of thinness was significantly lower in those who consumed at least 75% of the distributed sachets (150 micronutrient powder sachets) than in those who consumed less than 75% of the sachets (31% vs. 46%, p < .05). Anemia was prevalent among 56% and 55% of lactating mothers in the intervention and control areas, respectively. The hemoglobin level was higher and the anemia prevalence lower among those who consumed at least 75% of the distributed sachets than among those who consumed less (50% vs. 61%, p = .07) (table 4). In a multivariate analysis, lactating mothers who consumed at least 75% of the micronutrient powder sachets had 18% lower odds of being anemic than those in the control area, after adjustment for drinking water source and intake of green leafy vegetables, although the difference was not significant (OR = 0.82; 95% CI, 0.57 to 1.17) (table 5). On the other hand, lactating mothers who consumed less than 75% of the sachets had slightly higher odds of being anemic than those in the control area (OR = 1.17; 95% CI, 0.77 to 1.77). The mean (± SD) number of sachets consumed by the lactating mothers in the intervention area was 149 ± 61, resulting in a mean adherence rate of 75%; approximately 60% reported they had consumed more than 150 sachets. Adolescent girls

The mean (± SD) age of the adolescent girls in the intervention and control areas was 14.4 ± 1.1 and 14.6 ± 1.1 years, respectively. The mean weight and height were comparable in adolescent girls in the two areas (41.4 ± 5.9 vs. 41.3 ± 5.7 kg and 149.9 ± 5.6 vs. 149.8 ± 5.3 cm in the intervention and control areas, TABLE 4. Hemoglobin level and prevalence of anemia of lactating mothers in the intervention area according to rate of compliance with micronutrient powder use No. of sachets consumed Variable Hemoglobin level (g/dL) Anemia prevalence (%)

< 150 (n = 129)

≥ 150 (n = 190)

p

11.58 ± 1.49a

11.94 ± 1.43

.03

60.5

50.0

.07

a. Significantly different from hemoglobin level in women with higher compliance rate (p < .05, t-test).

283

Micronutrient powder use in Cyclone Sidr-affected areas

TABLE 5. Crude and adjusted odds ratios of the risk factors for anemia in lactating mothers n

Anemia (%)

Crude OR (95% CI)

Adjusted ORa (95% CI)

Intervention and compliance No intervention (control) Consumed < 150 micronutrient powder sachets Consumed ≥150 micronutrient powder sachets

361 129 190

55.4 60.5 50.0

1.0 1.23 (0.82–1.85) 0.81 (0.57–1.14)

1.0 1.17 (0.77–1.77) 0.82 (0.57–1.17)

Source of drinking water Not a tube well Tube well

94 624

67.0 54.0

1.0 0.58 (0.37–0.91)

1.0 0.62 (0.36–0.99)

Consumption of green leafy vegetables None Once a week or more

322 396

50.9 59.6

1.0 1.42 (1.06–1.91)

1.0 1.44 (1.06–1.96)

Factor

a. By multiple logistic regression analysis.

respectively). However, the prevalence of anemia was significantly higher among those in the intervention than among those in the control area (52% vs. 45%, p = .05), although neither group received micronutrient powder. Effects of Pushtika on food and side effects of Pushtika on beneficiaries

Among the lactating mothers and mothers of children under 5 years of age in the intervention area, 56% reported that Pushtika had had some kind of effect on the food to which it was added in the previous month. Approximately 42% of the mothers reported that the color of the food was affected, while a few reported that the texture (11%), smell (7%), or taste (10%) of the food was affected. However, of those who reported that Pushtika had had an effect on the food, only 12% replied that these effects had discouraged the use of Pushtika. Side effects associated with Pushtika consumption were rare. Only 1.7% of the mothers reported that they or their child had experienced side effects caused by Pushtika. The reported symptoms included nausea (0.7%) and vomiting (1.1%).

Discussion In this cross-sectional assessment of children under 5 years of age and lactating mothers severely affected by Cyclone Sidr, we examined whether the outcomes in the program areas in which micronutrient powder was provided alongside the general food rations in the emergency operation were different after 6 months of implementation from those in the nonprogram areas where micronutrient powder was not distributed. Although the associations found are potentially suggestive of impact, a causal association cannot be established in this study due to the cross-sectional design of the survey.

Hemoglobin level and anemia prevalence

The prevalence of anemia was alarmingly high among this emergency-affected population, with approximately 80% of children under 5 years of age and 55% of lactating mothers being anemic. The micronutrient powder was reported to be well accepted by the young children, with 78% having consumed more than 75% of the distributed sachets. Despite the high use of micronutrient powder, however, there were no differences in hemoglobin levels and anemia prevalence between children who received micronutrient powder and those who did not. Similarly, the prevalence of anemia did not differ between lactating mothers who received micronutrient powder and those who did not. However, these results need careful interpretation. Notably, among lactating women, the hemoglobin level was higher and anemia prevalence lower among those who consumed more than 75% of the distributed sachets than among those who consumed less. This suggests that micronutrient powder may positively affect the anemia prevalence of lactating mothers, provided that the compliance rate is high. Interestingly, it was also revealed that anemia prevalence was higher among adolescent girls in the intervention area than among those in the control area. Neither group of adolescents received micronutrient powder. This suggests that anemia prevalence prior to distribution of the micronutrient powder could also have been higher among children and lactating mothers in the intervention area than among those in the control area, in which case the micronutrient powder may have reduced anemia prevalence to the level observed in the control area. It is noteworthy that the anemia prevalence in both areas after the completion of the micronutrient powder program was still extremely high, indicating that even if the micronutrient powder had reduced anemia levels to the same level as those in the control area, the magnitude of reduction was not enough. This suggests that the children may be in need of more

284

micronutrient powder, for example, daily instead of every other day, and for a longer period of time. In addition, in reality the micronutrient powder was not distributed along with food as it had originally been planned (general food rations stopped as micronutrient powder started being distributed), and thus the content of certain micronutrients, such as vitamin A, might have been insufficient to meet the needs of the target beneficiaries. The efficacy of micronutrient powder has been widely examined, mostly in controlled studies among children [8]. A recent meta-analysis of home fortification of complementary foods by Dewey et al. found that in-home fortification with Sprinkles effectively reduced iron deficiency and decreased anemia prevalence by half [9]. Although the specific micronutrient composition is slightly different in Pushtika and Sprinkles, the formulation of Pushtika is composed of efficacious compounds based on current scientific knowledge of the best chemical forms and their interactions. Thus, the efficacy of Pushtika for the treatment and prevention of anemia can be assumed. In spite of the apparent efficacy of micronutrient powder, the present assessment failed to show that consumption of Pushtika by children in this particular emergency context was effective in reducing anemia. Alternatively, the prevalence of anemia might have been higher among children in the intervention than among those in the control area prior to micronutrient powder distribution and could have subsequently been reduced by the micronutrient powder to a level comparable to that in the comparison group. Compliance with micronutrient powder use was determined on the basis of self-report and recall by mothers, which may not reflect the true consumption behavior of the target beneficiaries. Thus, it cannot be ruled out that the true compliance rate might be lower than reported. A positive impact of micronutrient powder on hemoglobin levels and anemia prevalence among lactating mothers was only observed among those who reported a high intake of micronutrient powder. This is expected, since a good compliance (i.e., consumption of the product) is a prerequisite for the micronutrient powder product to have an impact. Anthropometric measurements

Child undernutrition was widespread in this population, with more than 40% of children under 5 years of age being stunted and 12% being wasted. There were no differences in the prevalence of undernutrition between children who received micronutrient powder and those who did not. This may suggest that micronutrient powder intervention was not effective in improving child growth and nutritional status, based on the assumption that the height and weight were

J.-H. Rah et al.

comparable between the two groups prior to distribution of the micronutrient powder. Our findings are consistent with the conclusion drawn from the metaanalysis of Dewey et al. that home fortification with micronutrients alone has no impact on child growth [9]. Nonetheless, it is noteworthy that the prevalence of stunting was lower among children who reportedly consumed more than 75% of the distributed sachets than among those who consumed less. Similarly, maternal thinness, indicating chronic energy deficiency or potentially acute malnutrition due to Cyclone Sidr, was less prevalent among lactating women who had a higher intake of micronutrient powder. These findings indicate that micronutrient powder consumption may be protective against child stunting and maternal thinness, provided that compliance is high. Alternatively, the use of micronutrient powder may have resulted in modifications to child feeding practices, which may have been the actual cause of impact on stunting. Morbidity

We found no differences between groups in morbidity symptoms among children and lactating mothers. Most previous efficacy studies found mixed results with regard to the impact of micronutrient powder on morbidity, including diarrhea, respiratory illness, and fever. The meta-analysis of Dewey et al. concluded that micronutrient powder consumption had a beneficial impact on morbidity in some high-risk populations, but most studies showed no significant impact [9]. Lessons learned—evaluation of programs

It is worth emphasizing the challenges of implementing a micronutrient powder program at scale and designing a proper evaluation in an emergency context, where a baseline assessment is often impossible because of constraints in time and expertise at hand and the impossibility of withholding micronutrient powder from the beneficiaries while all the preparatory activities for the baseline assessment take place. Thus, as long as the effectiveness of MNP under program conditions are proven, in an emergency context, it may be advisable to focus on program implementation and assessment of acceptance and adherence rather than on a properly designed impact evaluation. For future micronutrient powder programs implemented in development settings, a carefully designed impact assessment is needed to enhance our understanding of the impact of micronutrient powder on the health and nutritional status and micronutrient deficiencies among the vulnerable segments of the population. Future surveys and assessments aiming to gauge the effectiveness of micronutrient powder intervention should be carefully designed, taking into consideration the necessity of having a valid comparison group and

285

Micronutrient powder use in Cyclone Sidr-affected areas

baseline measurements taken prior to micronutrient powder distribution, preferably from the same group of people. In addition, efforts need to be made to develop and establish a monitoring system through which the actual number of micronutrient powder sachets distributed and consumed by the beneficiaries can be tracked at the household level.

Acknowledgments The present program was supported by DSM (a global Life Sciences and Materials Sciences Company addressing some of the world’s most pressing issues, such as

climate change, energy consumption, and the need for a balanced food supply) as part of a joint initiative by the WFP–DSM partnership known as “Improving Nutrition—Improving Lives.” The micronutrient powder was also produced and donated by DSM. The authors wish to acknowledge Joris van Hees, WFP, Rome, for his input throughout the program and Jonathan Sugimoto for his assistance and guidance in the statistical analysis. Additional appreciation goes to Rezaul Karim and Britta Schumacher at the Bangladesh Country Office of WFP for their comments and guidance throughout the program. We would also like to thank all the field staff of Nielsen Company in Bangladesh for their work and dedication during the data collection.

References 1. World Health Organization/World Food Programme/ United Nations Children’s Fund. Preventing and controlling micronutrient deficiencies in populations affected by an emergency. Multiple vitamin and mineral supplements for pregnant and lactating women, and for children aged 6 to 59 months. Geneva: WHO/WFP/ UNICEF, 2007. 2. World Health Organization/Food and Agriculture Organization. Vitamin and mineral requirements in human nutrition, 2nd ed. Geneva: WHO/FAO, 2004. 3. Dewey KG, Adu-Afarwuah S. Systematic review of the efficacy and effectiveness of complementary feeding interventions in developing countries. Matern Child Nutr 2008;4(suppl):24-85. 4. Menon P, Ruel MT, Loechl CU, Arimond M, Habicht JP, Pelto G, Michaud L. Micronutrient Sprinkles reduce anemia among 9- to 24-mo-old children when delivered through an integrated health and nutrition program in rural Haiti. J Nutr 2007;137:1023-30. 5. de Pee S, Moench-Pfanner R, Martini E, Zlotkin SH, Darnton-Hill I, Bloem MW. Home fortification in emergency response and transition programming:

6.

7. 8. 9. 10.

Experiences in Aceh and Nias, Indonesia. Food Nutr Bull 2007;28:189-97. de Pee S, Kraemer K, van den Briel T, Boy E, Grasset C, Moench-Pfanner R, Zlotkin S, Bloem MW. Quality criteria for micronutrient powder products: Report of a meeting organized by the World Food Programme and Sprinkles Global Health Initiative. Food Nutr Bull 2008; 29:232-41. Gibson RS. Principles of nutritional assessment. New York: Oxford University Press, 1990. Zlotkin SH, Schauer C, Christofides A, Sharieff W, Tondeur MC, Hyder SM. Micronutrient Sprinkles to control childhood anemia. PloS Med 2005;2:e1:24-8. Dewey KG, Yang Z, Boy E. Systematic review and metaanalysis of home-fortification of complementary foods. Matern Child Nutr 2009;5:283-321. WHO Multicentre Growth Reference Study Group (2006) WHO Child Growth Standards: Length/heightfor-age, weight-for-age, weight-for-length, weight-forheight and body mass index-for-age: Methods and development. Geneva: WHO.

Relationship of the availability of micronutrient powder with iron status and hemoglobin among women and children in the Kakuma Refugee Camp, Kenya Philip Ndemwa, Christine L. Klotz, David Mwaniki, Kai Sun, Erastus Muniu, Pauline Andango, Joyce Owigar, Jee Hyun Rah, Klaus Kraemer, Paul B. Spiegel, Martin W. Bloem, Saskia de Pee, and Richard D. Semba Abstract Background. Micronutrient powder is a potential strategy to improve iron status and reduce anemia in refugee populations. Objective. To evaluate the effect of the availability of home fortification with a micronutrient powder containing 2.5 mg of sodium iron ethylenediaminetetraacetate (NaFeEDTA) on iron status and hemoglobin in women and children in the Kakuma Refugee Camp in northwest Kenya. Methods. Hemoglobin and soluble transferrin receptor were measured in 410 children 6 to 59 months of age and 458 women of childbearing age at baseline (just before micronutrient powder was distributed, along with the regular food ration) and at midline (6 months) and endline (13 months) follow-up visits. Results. At the baseline, midline, and endline visits, respectively, the mean (± SE) hemoglobin concentration in women was 121.4 ± 0.8, 120.8 ± 0.9, and 120.6 ± 1.0 g/L (p = .42); the prevalence of anemia (hemoglobin < 120 g/L) was 42.6%, 41.3%, and 41.7% (p = .92); and the mean soluble transferrin receptor concentration was 24.1 ± 0.5, 20.7 ± 0.7, and 20.8 ± 0.7 nmol/L (p = .0006). In children, the mean hemoglobin concentration was 105.7 ± 0.6, 109.0 ± 1.5, and 105.5 ± 0.3 g/L (p = .95), respectively; the prevalence of anemia (hemoglobin < 110 g/L) was 55.5%, 52.3%, Philip Ndemwa, David Mwaniki, Erastus Muniu, and Pauline Andango are affiliated with the Kenya Medical Research Institute, Nairobi. Christine L. Klotz, Joyce Owigar, Martin W. Bloem, and Saskia de Pee are affiliated with the World Food Programme, Rome, Italy. Kai Sun, Jee Hyun Rah, and Klaus Kraemer are affiliated with Sight and Life, Basel, Switzerland. Richard D. Semba is affiliated with the Johns Hopkins University School of Medicine, Baltimore, Maryland, USA. Paul B. Spiegel is affiliated with the United Nations High Commissioner for Refugees, Geneva, Switzerland. Please direct queries to the corresponding author: Richard D. Semba, Johns Hopkins University School of Medicine, Smith Building, M015, 400 N. Broadway, Baltimore, MD 21287, USA; e-mail: [email protected].

286

and 59.8% (p = .26); and the mean soluble transferrin receptor concentration was 36.1 ± 0.7, 29.5 ± 1.9, and 28.4 ± 3.2 nmol/L (p = .02), in models that were adjusted for age using least squares means regression. Conclusions. In children and in women of childbearing age, the availability of micronutrient powder was associated with a small improvement in iron status but no significant change in hemoglobin in this refugee camp setting.

Key words: Anemia, children, hemoglobin, iron, micronutrient powder, refugees, women

Introduction Anemia and iron-deficiency anemia are extremely common among women and children in sub-Saharan Africa [1]. Anemia, as one indicator of micronutrient deficiencies, is largely attributed to iron deficiency, but it is also related to deficiencies of vitamin A, folic acid, and vitamin B12 [2]. In addition, non-nutritional factors such as parasitic infections, malaria, infectious diseases, and hemoglobinopathies contribute to anemia [3]. The main cause of micronutrient deficiencies in developing countries is a poor-quality diet that is largely plant-based, with limited inclusion of animalsource foods and fortified food products [4, 5]. Micronutrient malnutrition is common among people who consume a largely plant-based diet that also includes few fortified foods. In diets that are high in unrefined grains and legumes, the amount of nutrients consumed may be adequate, but dietary constituents, such as phytates and tannins, limit their absorption [6]. Home fortification with micronutrient powder has been developed to ensure an adequate intake of micronutrients in specific, high-risk target populations. Micronutrient powder is used to fortify ready-to-eat foods, such as porridge or a solid meal, at home just before consumption [7]. The powder is generally dosed to provide one full Recommended Nutrient Intake

Food and Nutrition Bulletin, vol. 32, no. 3 © 2011, The United Nations University.

Availability of micronutrient powder in relationship to iron status and hemoglobin

(RNI) of each of several vitamins and minerals in the form of 1 g of powder contained in a sachet. The formulation may vary according to local conditions, such as the availability of fortified food rations and the demographic characteristics of the target group. Micronutrient powder has been shown to reduce anemia and iron deficiency among young children in various settings [8]. Experience with its use in largescale programs has been gathered more recently, mainly in development settings, but also in a few emergency and refugee settings, including in Aceh, Indonesia [9]. The Kakuma Refugee Camp is located in Turkana District in the Rift Valley Province, about 110 km from the Sudanese border at Lokichoggio and about 50 km from the Ugandan border. The altitude of the camp is 580 m. The camp was established in 1992 to accommodate about 17,000 southern Sudanese refugees fleeing the civil war in their country. In 2009, the Kakuma Refugee Camp had a population of approximately 51,000 people, mostly Sudanese, Somalis, and Ethiopians. The refugees in the camp are provided with insecticide-treated mosquito nets and have access to medical treatment in the International Red Cross facilities. The population of the camp relies upon the general food ration distributed by the United Nations World Food Programme (WFP). The food ration consists of maize, fortified wheat flour, pulses, vegetable oil fortified with vitamin A, fortified corn–soy blend, and iodized salt. The refugees also generally receive complementary foods such as green grams or groundnuts from the United Nations High Commissioner for Refugees. The refugees have limited access to fresh foods. The food ration provides an average of 2,100 kcal/ person/day, but it does not contain all the required micronutrients. For the average person, based upon the specific age and sex distribution of the population, the general food ration provides 81% of iron, 43% of vitamin C, 62% of riboflavin, and 89% of thiamin requirements. Moreover, micronutrient intake is further reduced by poor bioavailability of some of the micronutrients, particularly iron and zinc. Whether home fortification with micronutrient powder can reduce anemia in large refugee camp settings in sub-Saharan Africa has not been well characterized. We hypothesized that the availability of micronutrient powder for home fortification would be associated with an improvement in iron status and hemoglobin concentrations in women and preschool-aged children in the Kakuma Refugee Camp in northwest Kenya. Because the camp is located in a malaria-endemic area, the quantity of iron in the micronutrient powder was reduced compared with the micronutrient powder used in areas where there is little to no malaria. The form of iron in the powder was sodium iron ethylenediaminetetraacetate (NaFeEDTA), which has high bioavailability. To address the hypothesis, we measured iron status

287

using serum soluble transferrin receptor, a marker for tissue iron deficiency [10], and hemoglobin concentrations among women of childbearing age and children aged 6 to 59 months, before and after micronutrient powder was made available in food distribution centers within the camp.

Materials and methods The study population consisted of 410 children 6 to 59 months of age and 458 women of childbearing age (18 to 49 years) in the Kakuma Refugee Camp in northwest Kenya. Subjects were eligible for the study if they were resident in the camp with no plans to move in the next 12 months, were children aged 6 to 59 months or nonpregnant women of reproductive age (18 to 49 years), and agreed to participate in the study and give written, informed consent or had a parent or guardian who gave written, informed consent. Subjects were excluded from the study and referred to the health center for further management if they had a hemoglobin level < 80 g/L at the time of screening. The participants in the study were a representative sample of the population in the Kakuma Refugee Camp. Households were selected for the study by cluster sampling (probability proportional to size) of 32 households from 25 administrative units in the camp. Subjects were enrolled in January 2009 in a baseline survey. The cohort was seen at midline and endline follow-up visits at 6 and 13 months after the baseline survey. The study protocol was approved by the Institutional Review Boards of both the Johns Hopkins University School of Medicine and the Kenya Medical Research Institute. Boxes containing 30 sachets of MixMe micronutrient powder (DSM Nutritional Products Ltd., Kaiseraugst, Switzerland) were offered at the monthly food distribution at the camp beginning in the month following the baseline study visit. Prior to and in the first months of distribution of the micronutrient powder, social marketing efforts were undertaken to inform the inhabitants of the camp about the potential health benefits and proper use of the micronutrient powder. The micronutrient formulation was specifically tailored for the Kakuma Refugee Camp setting. The refugees were educated to use one sachet per family member each day, mixing the contents with food that was ready for consumption. One sachet of the micronutrient powder (1 g ) contained vitamin A (100 μg RE), vitamin D3 (5 μg), vitamin E (5 mg), vitamin K1 (30 μg), thiamin (0.5 mg), riboflavin (0.5 mg), pyridoxine (0.5 mg), folic acid (90 μg), niacin (6 mg), vitamin B12 (0.9 μg), vitamin C (60 mg), iron (2.5 mg), zinc (2.5 mg), selenium (17 μg), copper (0.34 μg), iodine (30 μg), and maltodextrin as a carrier. Iron was calculated based upon reference values

288

for medium bioavailability. The iron content of the formulation was reduced from 10 mg per serving to 2.5 mg in order to minimize the risk due to use of iron in malaria-endemic areas [11]. The form of iron was NaFeEDTA, a highly bioavailable form. The iron content of 2.5 mg was meant to complement approximately 3 mg of iron that was provided by fortification iron in corn–soy blend and wheat flour. With total iron absorption from the fortified foods and the micronutrient powder of 10%, it was estimated that the 0.55 mg of iron absorbed would provide most of the daily requirement of a 1- to 3-year-old child. The zinc content was reduced to 2.5 mg to maintain an equal ratio to iron, as zinc can interfere with the absorption of iron. The proportion of families who picked up the boxes of micronutrient powder was monitored at each of the monthly food distribution cycles. A 4-day training and a 1-day pretest were conducted prior to each of the three surveys to equip the survey staff with the objectives of the micronutrient powder intervention and techniques for accurate collection of data (blood samples, anthropometric measurements, and administration of questionnaires). The teams were composed of local agency staff (community health workers and laboratory staff), national staff supervisors (health personnel and nutritionists), and data entry clerks. Nine teams of six members each (one supervisor, one enumerator, two anthropometrists, and two phlebotomists) collected the survey data at baseline, midline, and endline. Interviews were conducted in the language of the respondent and translated back into English for enumeration in a standardized questionnaire. The same survey staff members were used wherever possible, although many new workers were recruited for the midline and endline surveys due to the high turnover of local workers in the camp. The Principal Investigator from the Kenya Medical Research Institute (P.N.) led a quality control team to recheck anthropometric measurements. After the mother, father, or guardian gave written, informed consent, data regarding basic demographic characteristics, morbidity, food intake, food security, and knowledge of anemia were collected using standardized questionnaires. Children were weighed with UNICEF Salter scales. Children were minimally or lightly clothed and placed in the weighing pants, with weight recorded to the nearest 0.1 kg. Height was measured to the nearest 0.1 cm with wooden height boards. Children under 24 months of age were measured in the recumbent position; older children were measured in the upright position. Women were weighed to the nearest 0.1 kg with electronic scales. Adult height boards were used to measure heights of women. According to data from the United Nations High Commissioner on Refugees Health Information System, the crude malaria incidence per month (per 1,000 persons) during the 13-month study period from

P. Ndemwa et al.

January 2009 through February 2010 was 20.75, 27.76, 16.04, 32.47, 21.43, 23.93, 19.07, 16.14, 23.09, 27.69, 20.85, 23.17, 48.56, and 32.83, with no temporal trend by logistic regression (p = .13). The malaria incidence among children under 5 years of age per month (per 1,000 persons) during the same time period was 74.33, 87.42, 61.48, 154.27, 85.66, 95.35, 68.89, 66.61, 92.2, 107.84, 89.77, 85.28, 152.42, and 110.62 (p = .24). Fingerstick blood samples were collected using capillary tubes at each visit. Hemoglobin was measured by HemoCue. Capillary tubes were stored in cool boxes with gel packs for 2 to 4 hours until they were processed in a central laboratory. Serum was separated by centrifugation, aliquoted into cryovials, and immediately placed in liquid nitrogen refrigerators in the field. Samples were later transferred to a freezer for long-term storage at −70° C. Soluble transferrin receptor was measured in serum by a commercial ELISA (Human sTfR Quantikine, R & D Systems). In a coinvestigator’s laboratory (R.D.S.), the interassay and intraassay coefficients of variation were 5.3% and 4.7%, respectively. The lower limit of detection of the assay was 0.5 nmol/L, and the range of the assay was 3 to 80 nmol/L, according to the manufacturer’s description. No samples were below the limit of detectability or above the range of the assay. Statistical analysis

The sample size of the study was based upon a 90% power to detect at least a 20% reduction in anemia in both women and children with a two-sided test, α = .05, and an estimated loss to follow-up of 30%. Continuous variables were described by means and standard errors. Skewed variables were log transformed to achieve a normal distribution. Weight-for-age, height-for-age, and weight-for-height z-scores were calculated using the World Health Organization (WHO) child growth reference population [12]. Underweight, stunting, and wasting, respectively, were defined as weight-for-age, height-for-age, and weight-for-height z-score < −2. Body mass index (BMI) was calculated as the weight in kilograms divided by the square of the height in meters. Anemia was defined as hemoglobin < 110 g/L for children and < 120 g/L for women, according to WHO definitions [3]. In children, least squares linear regression models were used to examine respective changes in hemoglobin and soluble transferrin receptor over time, with adjustment for child age. All analyses were performed by SAS with a type I error of 0.05.

Results The demographic and anthropometric characteristics of the 458 women and 410 children at baseline are shown in table 1. Of the 458 women seen at baseline,

Availability of micronutrient powder in relationship to iron status and hemoglobin

TABLE 1. Characteristics of women and children at the baseline visita Characteristic

Valueb

Women (n = 458) Age—yr

29.4 ± 0.4

BMI (kg/m2)—% < 18.5 18.5–24.9 25.0–29.9 ≥ 30

59.0 20.3 14.5 6.2

Ethnic group—% Sudanese Somali Ethiopian Ugandan or Congolese

49.3 39.1 9.4 2.2

Children (n = 410) Male sex—%

50.7

Age (mo)—% 6–11 12–23 24–35 36–47 48–59

9.3 21.7 23.7 22.4 22.9

Weight-for-age z-score % with z-score < –2

–0.57 ± 0.06 8.7

Height-for-age z-score % with z-score < –2

–0.16 ± 0.08 15.3

Weight-for-height z-score % with z-score < –2

–0.66 ± 0.06 13.6

BMI, body mass index a. Data are missing for maternal BMI (12 subjects), ethnicity (3 subjects), and child anthropometric measurements (9 subjects). b. Plus–minus values are means ± SE.

175 (38.2%) were lost to follow-up by the endline visit. Of the 410 children seen at baseline, 111 (27.1%) were lost to follow-up by the endline visit. There were no significant differences in age, BMI, ethnicity, hemoglobin concentration, or soluble transferrin receptor concentration between women who were lost to follow-up and women who remained in the study (data not shown). There were no significant differences in age, sex ratio, weight-for-age, height-for-age, weight-for-height, hemoglobin concentration, or soluble transferrin receptor concentration between children who were lost to follow-up and children who remained in the study (data not shown). The proportion of boxes of micronutrient powder picked up at the monthly food distribution in the camp were as follows: February 2009, 99%; March 2009, 85%; April 2009, 71%; May 2009, 60%; June 2009, 49%; July 2009, 30%; August 2009, 45%; September 2009, 40%; October 2009, 48%; November 2009, 46%; December 2009, 47%; January 2010, 49%; and February 2010, 50%.

289

In women of childbearing age, the mean hemoglobin concentration and the prevalence of anemia did not change significantly from baseline through midline and endline (table 2). The mean soluble transferrin receptor concentration decreased between baseline and endline (p = .0006). In children, after adjustment for age by least squares means regression, the mean hemoglobin concentration did not change significantly from baseline through endline (table 2). The prevalence of anemia, adjusted for age by least squares means regression, did not change significantly between baseline and endline. After adjustment for age by least squares means regression, the mean soluble transferrin receptor concentration decreased between baseline and endline (p = .02).

Discussion The findings of the present study, although relatively modest, corroborate the findings of other studies that show home fortification using micronutrient powder improves iron status. The use of home fortification has consistently improved iron status of children in experimental studies [13–16], even in malaria-endemic areas. Although the present study was limited by the absence of a control group and cannot rule out other factors as being responsible for the observed improvement in iron status, the findings from other studies on home fortification and the efficacy of NaFeEDTA in high-phytate diets support an effect of the iron intervention [17]. Low compliance may have played a role in the results of the present study. The barriers to use of the micronutrient powder in the camp are presented in a separate paper [18]. A common factor in the studies cited above is a compliance rate that exceeded 80% [13, 16, 19]. In a study by Suchdev and colleagues in western Kenya [20], final hemoglobin concentrations were associated with the number of micronutrient powder sachets purchased. Given that availability was assessed in terms of sachets purchased, the likelihood that these were also consumed may be higher than in our study, where picking up micronutrient powder sachets at the distribution center may not necessarily have translated to compliance in terms of consumption. The micronutrient powder used in the present study should deliver an amount of absorbable iron comparable to that delivered by the formulations used in most of the studies noted above [21] and hence should sufficiently improve iron status if compliance is good. It is probable, therefore, that with improved compliance, home fortification could achieve greater gains in reducing the prevalence of anemia in refugee camps. The micronutrient powder was especially formulated for the situation of the Kakuma Refugee Camp and did not contain the Recommended Nutrient Intakes (RNIs) of iron and vitamin A. The iron content was reduced

290

because malaria is endemic in the region. Absorption of iron is expected to be more gradual with the chelated form of iron and less likely to cause a peak of non-transferrin-bound iron. The vitamin A content was reduced because vitamin A–fortified cooking oil is provided in the monthly food basket by WFP. The limitations of this study include the high rate of loss to follow-up and low uptake of the micronutrient powder. The high loss to follow-up reached nearly 40% for women of childbearing age and nearly 30% for preschool-aged children and is largely attributed to the population dynamics of the camp. The proportion of adults who collected the micronutrient nutrient powder dropped steadily from 99% to 30% within the first 6 months that the powder was available at the food distribution center. In the remaining 7 months of the study, the proportion of adults who collected the micronutrient powder at the center remained fairly steady at 40% to 50%. It is interesting to note that among women of childbearing age and children, the decrease in mean soluble transferrin receptor occurred from baseline to midline visits and then showed no change from midline to endline visits. Soluble transferrin receptor is a sensitive indicator of tissue iron deficiency and is less affected by inflammation than is serum ferritin [10]. Although there was a significant decrease in soluble transferrin receptor in women of childbearing age and children during the study, it appears that the improvement in iron status was not sufficient to increase hemoglobin concentrations or the prevalence of anemia during follow-up. The mean soluble transferrin receptor concentrations at midline and endline visits in women were close to the mean levels of 19.5 nmol/L reported in a study of 225 healthy white, black, and Hispanic adults and 20.9 nmol/L reported in 61 black adults [22]. Agespecific soluble transferrin receptor concentrations

P. Ndemwa et al.

in healthy children aged 2 to 6 years showed a reference interval (2.5th to 97.5th percentile) of 14.0 to 40.8 nmol/L [23]. The mean soluble transferrin receptor concentrations of children in the present study were toward the upper end of this reference interval. In conclusion, the availability of micronutrient powder was associated with a small improvement in iron status in children and women of childbearing age in the Kakuma Refugee Camp. Further work is needed to address and overcome the barriers to utilization of micronutrient powder in this refugee camp setting.

Acknowledgments The micronutrient powder program implementation and study at Kakuma Refugee Camp was carried out and funded as part of the WFP–DSM partnership “Improving Nutrition—Improving Lives” aiming to assist WFP in implementing its Nutrition Improvement Approach. The following individuals substantially contributed to the project: Sammy Nadiko, Simon Elimo, Daniel Macharia, and Joseph Kimani (monitoring team); Patrick Mundia, Joyce Mukiiri, and Carolyne Mbuuri (International Rescue Committee staff); Victoria Mwenda and Caroline Wilkinson (United Nations High Commissioner for Refugees staff ); Lourdes Ibarra, Peter Otieno, Josiah Osiri, Beatrice Ngugi, Charles Owade, Ronald Erukan, Jecinta Abenyo, Peter Karuri, Thomas Chika, Wilson Ereng, Alfred Lokwang, Daniel Mwangi, and David Kariuki (WFP Kakuma); Emily Madete, Grace Igweta, Stephen Mwenda, Felix Okech, and Josephine Mahiga (WFP Nairobi); Nils Grede (WFP Rome); Sam Okwara, Ezekiel Mukhaye, and Saidi Kisiwa (Kenya Medical Research Institute staff); and Martin Meme and Emily Teshome (survey consultants).

References 1. Kraemer K, Zimmermann MB, eds. Nutritional anemia. Basel, Switzerland: Sight and Life Press, 2007. 2. Allen LH. Causes of nutrition-related public health problems of preschool children: available diet. J Pediatr Gastroenterol Nutr 2006;43:S8–12. 3. World Health Organization. Iron deficiency anaemia: Assessment, prevention and control: a guide for programme managers. WHO, 2001. Available at: http:// www.who.int/nutrition/publications/micronutrients/ anaemia_iron_deficiency/WHO_NHD_01.3/en/index. html. 4. de Pee S. Need, efficacy and effectiveness of multiple vitamin/mineral supplements for young children and considerations for programs. In: Semba RD, Bloem MW, eds. Nutrition and health in developing countries, 2nd ed. Totowa, NJ, USA: Humana Press, 2008:793–830. 5. Zimmermann MB, Hurrell RF. Nutritional iron deficiency. Lancet 2007;370:511–20.

6. Sotelo A, González-Osnaya L, Sánchez-Chinchillas A, Trejo A. Role of oxate, phytate, tannins and cooking on iron bioavailability from foods commonly consumed in Mexico. Int J Food Sci Nutr 2010;61:29–39. 7. de Pee S, Kraemer K, van den Briel T, Boy E, Grasset C, Moench-Pfanner R, Zlotkin S, Bloem MW. Quality criteria for micronutrient powder products: report of a meeting organized by the World Food Programme and Sprinkles Global Health Initiative. Food Nutr Bull 2008;29:232–41. 8. Dewey KG, Yang Z, Boy E. Systematic review and metaanalysis of home fortification of complementary foods. Matern Child Nutr 2009;5:283–321 9. de Pee S, Moench-Pfanner R, Martini E, Zlotkin SH, Darnton-Hill I, Bloem MW. Home fortification in emergency response and transition programming: experiences in Aceh and Nias, Indonesia. Food Nutr Bull 2007;28:189–197.

Availability of micronutrient powder in relationship to iron status and hemoglobin

10. Beguin Y. Soluble transferrin receptor for the evaluation of erythropoiesis and iron status. Clin Chim Acta 2003;329:9–22. 11. Sazawal S, Black RE, Ramsan M, Chwaya HM, Stoltzfus RJ, Dutta A, Dhingra U, Kabole I, Deb S, Othman MK, Kabole FM. Effects of routine prophylactic supplementation with iron and folic acid on admission to hospital and mortality in preschool children in a high malaria transmission setting: community-based, randomised, placebo-controlled trial. Lancet 2006;367:133–43. 12. World Health Organization Multicentre Growth Reference Study Group. WHO child growth standards: Growth velocity based on weight, length and head circumference: methods and development. Geneva: WHO, 2009. 13. Adu-Afarwuah S, Lartey A, Brown KH, Zlotkin S, Briend A, Dewey KG. Home fortification of complementary foods with micronutrient supplements is well accepted and has positive effects on infant iron status in Ghana. Am J Clin Nutr 2008;87:929–38. 14. Christofides A, Asante KP, Schauer C, Sharieff W, Owusu-Agyei S, Zlotkin S. Multi-micronutrient sprinkles including a low dose of iron provided as microencapsulated ferrous fumarate improves haematologic indices in anaemic children: a randomized clinical trial. Matern Child Nutr 2006;2:169–80. 15. Giovannini M, Sala D, Usuelli M, Livio L, Francescato G, Braga M, Radaelli G, Riva E. Double-blind, placebocontrolled trial comparing effects of supplementation with two different combinations of micronutrients delivered as sprinkles on growth, anemia, and iron deficiency in Cambodian infants. J Pediatr Gastroenterol Nutr 2006;42:306–12. 16. Zlotkin S, Arthur P, Schauer C, Antwi KY, Yeung G, Piekarz A. Home-fortification with iron and zinc sprinkles or iron sprinkles alone successfully treats anemia in

291

infants and young children. J Nutr 2003;133:1075–80. 17. Andang’o PE, Osendarp SJ, Ayah R, West CE, Mwaniki DL, De Wolf CA, Kraaijenhagen R, Kok FJ, Verhoef H. Efficacy of iron-fortified whole maize flour on iron status of schoolchildren in Kenya: a randomised controlled trial. Lancet 2007;369:1799–806. 18. Kodish S, Rah JH, Kramer K, de Pee S, Gittelsohn J. Understanding low usage of micronutrient powder in the Kakuma Refugee Camp, Kenya: Findings from a qualitative study. Food Nutr Bull 2011;32:291–302. 19. Zlotkin S, Antwi KY, Schauer C, Yeung G. Use of microencapsulated iron(II) fumarate sprinkles to prevent recurrence of anaemia in infants and young children at high risk. Bull World Health Organ 2003;81:108–15. 20. Suchdev P, Ruth L, Mandava U, Quick R, Mbakaya C, Kaduka L, Jefferds ME, Woodruff BA. Effectiveness of sprinkles sales in Western Kenya in reducing childhood anemia and iron deficiency [poster]. In: Proceedings of the Micronutrient Forum Meeting, 12–15 May, 2009, Beijing. Available at: http://www.micronutrientforum. org/Meeting2009/PDFs/Poster%20Presentations/3_ Thursday/SNT/TH41_Suchdev.pdf. Accessed 13 May 2011. 21. Verhoef H, Veenemans J. Safety of iron-fortified foods in malaria-endemic areas. Am J Clin Nutr 2009;89:1949–50. 22. Allen J, Backstrom KR, Cooper JA, Cooper MC, Detwiler TC, Essex DW, Fritz RP, Means RT Jr, Meier PB, Pearlman SR, Roitman-Johnson B, Seligman PA. Measurement of soluble transferrin receptor in serum of healthy adults. Clin Chem 1998;44:35–9. 23. Kratovil T, DeBerardinis J, Gallagher N, Luban NLC, Soldin SJ, Wong ECC. Age specific reference intervals for soluble transferrin receptor (sTfR). Clin Chim Acta 2007;380:222–4.

Understanding low usage of micronutrient powder in the Kakuma Refugee Camp, Kenya: Findings from a qualitative study

Stephen Kodish, Jee Hyun Rah, Klaus Kraemer, Saskia de Pee, and Joel Gittelsohn Abstract Background. Home fortification with micronutrient powder has been shown to be a low-cost, feasible, and effective approach to address micronutrient deficiencies. A large-scale program distributing micronutrient powder to approximately 50,000 refugees was implemented at the Kakuma Refugee Camp in Kenya. Uptake of the micronutrient powder at distribution points dropped nearly 70%, from 99% to a low of 30%, and remained at 45% to 52% despite increased social marketing efforts. Objective. To identify factors at the distal and proximal levels leading to the low uptake of micronutrient powder through a qualitative inquiry. Methods. In-depth interviews were conducted with community leaders, stakeholders, implementing partners, and beneficiaries. Direct observations of food preparation and child feeding were conducted. Focus group discussions were employed to examine perceptions and practices of beneficiaries regarding micronutrient powder use. Results. Superficial formative research and lack of interagency coordination led to insufficient social marketing prior to the program. In addition, community health workers were inadequately trained. This resulted in inadequate communication regarding the health benefits and use of micronutrient powder to the beneficiaries. Reliance on personal experiences with micronutrient powder and issues with its packaging, in part, led to confusion and deleterious rumors, resulting in decreased uptake of Stephen Kodish and Joel Gittelsohn are affiliated with the Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA; Jee Hyun Rah and Klaus Kraemer are affiliated with Sight and Life, Basel, Switzerland; Saskia de Pee is affiliated with the World Food Programme, Rome, and the Friedman School of Nutrition Science and Policy, Tufts University, Boston, Massachusetts, USA. Please direct queries to the corresponding author: Joel Gittelsohn, Center for Human Nutrition, Department of International Health, Johns Hopkins Bloomberg School of Public Health, Room W2041A, 615 N. Wolfe Street, Baltimore, MD, USA; e-mail: [email protected].

292

micronutrient powder at distribution points. Conclusions. A successful micronutrient powder program requires careful design, with emphasis on conducting thorough formative research, ensuring the involvement and commitment of all stakeholders from the outset, investigating the role of cultural factors, and ensuring provision of sufficient, adequate, and timely information to the beneficiaries.

Key words: Formative research, Kenya, micronutrient

deficiencies, micronutrient powder (MNP), program implementation, qualitative research, refugee camp, social marketing

Introduction Micronutrient malnutrition is an important public health problem, including among refugees dependent on food assistance. Limited access to fresh fruits and vegetables, restrictions on gardening and animal husbandry, and recurrent infectious disease due to poor sanitation and crowded living conditions increase the risk of micronutrient deficiencies and their consequences in protracted refugee settings [1]. Various attempts have been made to improve refugees’ micronutrient status through fortification of foods included in the general food ration and supplementation in the form of tablets, but with limited success. Although a sustainable, diverse diet is generally the most preferred approach, this is often not feasible due to the harsh climate and the shortage of cultivable land and resources available to the refugees [2]. In recent years, home fortification with micronutrients provided in the form of powders, crushable tablets, and lipid-based spreads has been recognized as a promising approach to prevent micronutrient deficiencies in vulnerable population groups, such as young children who cannot swallow tablets. In particular, the use of micronutrient powder, a mix of vitamins and minerals in powder form that can be added to food just before

Food and Nutrition Bulletin, vol. 32, no. 3 © 2011, The United Nations University.

293

Qualitative inquiry of micronutrient powder program

consumption, has gained popularity since its inception in the late 1990s due to ease of use and low cost [3]. There has also been abundant scientific evidence from diverse settings demonstrating the efficacy of micronutrient powder in the treatment and prevention of anemia, primarily among young children [4–7]. In general, micronutrient powder is packaged in singledose sachets targeted to provide 1 RNI (Recommended Nutrient Intake) of micronutrients per person per day. In continuing efforts by the United Nations World Food Programme (WFP) and the United Nations High Commissioner for Refugees (UNHCR) to address the high prevalence of undernutrition and micronutrient deficiencies (assessed by anemia) among refugees, a large-scale micronutrient powder program targeting the entire population of the Kakuma Refugee Camp in Kenya was initiated in February 2009. This was enabled by the partnership of WFP and DSM* known as “Improving Nutrition—Improving Lives.” The objectives of the micronutrient powder program were to reduce the prevalence of micronutrient malnutrition among the refugees by the 17-month provision of home fortification with micronutrient powder, and to determine the feasibility of distributing micronutrient powder in combination with general food rations in a refugee camp setting. The program provided each individual with a once-a-day micronutrient powder sachet with the brand name MixMe (DSM Nutritional Products Ltd., Kaiseraugst, Switzerland) containing a low dose of iron (2.5 mg of sodium iron ethylenediaminetetraacetate [NaFeEDTA]) and 15 other vitamins and minerals for a period of 17 months. One box containing 30 micronutrient powder sachets for each family member** was distributed monthly, together with standard food rations. Multiple monitoring assessments soon indicated that the rate of uptake*** of the micronutrient powder at the distribution points was dropping steadily by approximately 10% each month, from a high of 99% in February 2009 to a low of 30% in July 2009. Social marketing efforts, including a house-to-house campaign, were increased, which helped improve uptake by 15% by August 2009 to reach a steady collection rate of * DSM is a multinational life sciences and material sciences company and the world’s leading supplier of vitamins, carotenoids, and other health ingredients. ** To our knowledge, this is the first program that provided micronutrient powder to everyone aged 6 months and older, instead of only to a specific target group, such as children under 5 years of age. The reason this was done was that the implementing partners in the Kakuma Refugee Camp had advised that because of the practice of shared-plate feeding, it would not be feasible to add the micronutrient powder only to the meals of specific individuals in a household. *** The word “uptake” refers to the collection of micronutrient powder at food distribution centers. Beneficiaries were given the choice of picking up the micronutrient powder as they passed through the food distribution line.

approximately 45% to 52% until June 2010, when the program came to an end. However, anecdotal reports indicated that many refugees who collected the sachets at the distribution point soon discarded them. In this context, a qualitative inquiry was conducted to help understand the underlying causes of low micronutrient powder acceptance at the Kakuma Refugee Camp by evaluating the challenges and enabling factors for implementation of the program and to make recommendations for future micronutrient powder programs in Kenya and elsewhere. The research had two overarching aims: to understand perceptions of micronutrient powder and associated underlying causes of low uptake, and to make recommendations on how to improve micronutrient powder use both at Kakuma and at future intervention sites.

Methods Setting

The Kakuma Refugee Camp is located in the Turkana District in the northwestern part of Kenya. The area is characterized by semiarid climatic conditions with daily average temperatures of approximately 40°C. The camp hosts a fluctuating population of more than 50,000 refugees whose food security and health-related needs are primarily addressed by WFP and UNHCR. The population is diverse, with more than 10 ethnic groups from Somalia, Sudan, Ethiopia, Democratic Republic of Congo, and other countries. Numerous languages and dialects are spoken within the camp. At the time of this study, approximately 70% of the camp population was Somali. The movement of the refugees is restricted mainly to within the camp. In addition, because of the Kenyan encampment policy, refugees are restricted from earning full wages or raising livestock or crops; therefore, they depend almost entirely on a general food ration and supplementary foods for special target groups provided by WFP and limited fresh foods and special foods from UNHCR. The general food ration commodities, such as fortified blended food (corn–soy blend), wheat flour, peas, fortified vegetable oil, and iodized salt are distributed at the Kakuma food distribution centers twice a month. Micronutrient powders were distributed only during the first cycle of the twice-monthly food distributions. The whole camp population aged 6 months and above was targeted. Each individual was provided with enough micronutrient sachets to allow the consumption of one sachet each day. In March 2008, prior to the initiation of the micronutrient powder program, a 5-week acceptability study was undertaken at the camp to determine the feasibility of a micronutrient powder initiative and to test the acceptability of a potential product. Based on

294

the feedback from the feasibility study, which indicated potential widespread acceptability of the product, permission was sought from the Kenyan government to implement the proposed micronutrient powder intervention at the Kakuma Refugee Camp. Social marketing campaigns prior to and over the course of the program were undertaken to sensitize the refugees to MixMe. A cohort study conducted by the Johns Hopkins University School of Medicine in collaboration with the Kenya Medical Research Institute investigating the impact of micronutrient powder on anemia prevalence and nutritional status of young children and women was nested into the program, the results of which are reported in a separate paper in the present issue. The micronutrient powder program was initially planned to begin in the fall of 2008, but obtaining ethical clearance from the Kenya Medical Research Institute and the Johns Hopkins University School of Medicine for the cohort study delayed the launch of the program until February 2009. The program began with a high uptake, reflected by the high collection rate of 99% at the food distribution center in February 2009, which ultimately came down to an average uptake of about 45% to 50% toward the completion of the program. Qualitative research approach

The present qualitative research was carried out from July to September 2010, after the completion of the micronutrient powder program in June of the same year. Qualitative research uses an exploratory and iterative approach, often referred to as emergent design. The research was conducted in four phases, with each phase building upon information collected in the previous phase. Qualitative research seeks to elicit contextually rich results that are understandable and credible, both to the people being studied and to others [8]. Triangulation of methods, theories, and investigator perspectives is used to build context and complete understandings of the phenomenon under consideration [9]. Our study used direct observations, in-depth interviews, and focus group discussions (table 1). In addition, various cognitive anthropological methods—free lists, pile sorts, and paired comparisons (table 2)—provided complementary data [10]. For example, during free listing exercises, the participants were asked to list as many items in a specific cultural domain as they could (e.g., illnesses in the community). Follow-up questioning about the lists provided textual data, and frequencies and measures of saliency were derived through quantitative analyses of the items in the lists [10]. This paper discusses only data from in-depth interviews, focus group discussions, and free lists. A total of 10 data collectors were recruited from the refugee camp population to collect data from adult Somali, Somali Bantu, Congolese, Ethiopian, Sudanese Dinka, and Sudanese Nubian, the six largest

S. Kodish et al.

communities at the camp. Because of the heterogeneity of the camp population, data were collected separately in each ethnic community, a research design that allowed for within- and between-group analyses. Each data collector had completed a secondary school education and could read and write English. They collected data from individuals only in their own respective ethnic groups. After the recruitment, the data collectors were trained in an intensive 2-week qualitative training course by the principal investigators and received over 60 hours of classroom and field practice in qualitative methods before formal data collection commenced. The study protocol was approved by the Institutional Review Board of the Johns Hopkins School of Public Health and the Kenya Council for Science. Oral informed consent was obtained from all participants. Phase 1: In-depth interviews with community leaders

In-depth interviews and free-listing activities with 20 community leaders were employed to identify important local health concerns, provide an overview of primary issues and barriers associated with micronutrient powder introduction and family feeding practices, and develop beneficiary in-depth interview questions. Leader in-depth interviews also helped to identify positive deviants among micronutrient powder beneficiaries for phase 2 interviewing [11]. Phase 2: Direct observations and in-depth interviews with micronutrient powder program beneficiaries

The second phase sought to understand the context and patterns of in-home eating behavior of refugees, with an emphasis on food preparation and feeding of young children. The data collectors observed meal preparation and feeding behavior and wrote detailed notes using a semistructured form. Twenty-one direct observations were conducted. In addition, the data collectors conducted 45 TABLE 1. Summary of data collection methods Method

No. conducted

Type of participant

In-depth interviews

45 20 14

Beneficiary Community leader Stakeholder and implementing partner

Focus group discussions

10

Beneficiary

Direct observations

21

Beneficiary

Free lists

72

Community leader and beneficiary

Pile sorts

34

Beneficiary

Paired comparisons

70

Beneficiary

295

Qualitative inquiry of micronutrient powder program

TABLE 2. Cultural domain analysis techniques and usage Structured interviewing method Description

Application in the current study

Free lists

Simple listing activity to identify important items in identified cultural domains

Community leaders were asked to list common illnesses and the effective treatments for anemia in their communities

Pile sorts

Sorting activity used to determine how items within an identified domain are related to one another

Beneficiaries freely sorted the most salient illnesses and the treatments for anemia in their communities; interview questions followed

Paired comparisons

Ranking activity for finding the relative importance of cultural domain items to a person or group

Beneficiaries ranked salient illnesses based on severity and anemia treatments on perceived effectiveness; interview questions followed

in-depth interviews with the beneficiaries about general eating behavior, factors related to micronutrient powder use, strategies for micronutrient powder adherence, and recommendations for program improvement. Data were collected with the use of semistructured interview guides that allowed the data collectors to probe on areas of interest related to the research questions and aims. These respondents were purposively sampled within each of the six communities, with the help of the community leaders, to represent a range of experiences with the micronutrient powder project. Of particular interest were mothers with young children and individuals who used the micronutrient powder despite its challenges. All in-depth interviews lasted approximately 45 to 60 minutes and were conducted until all primary and tertiary factors that led to low micronutrient powder acceptance were elucidated within each community. Phase 3: In-depth interviews with stakeholders and implementing partners

In this phase, 14 individuals identified as the key stakeholders of the Kakuma micronutrient powder program and those who played otherwise crucial roles in program implementation were interviewed about their experiences with the program and their perspectives on reasons for its challenges. These interviews were conducted in English by the primary investigators (Stephen Kodish and Joel Gittelsohn) with key individuals from the collaborating organizations, including DSM, WFP, UNHCR, the Lutheran World Federation (LWF), FilmAid, the International Rescue Committee (IRC), and Population Services International (PSI). Phase 4: Focus group discussions

Data, including emergent themes, from in-depth interviews, pile sorts, and free lists were used to develop potential strategies for modifying the micronutrient powder product itself for future programs. Ten ethnically homogeneous focus group discussions were conducted. Each group was composed of 6 to 10 persons, and the discussion lasted about 1.5 hours. The

groups were either all female or all male. Research team members moderated the sessions in local languages to identify social norms related to causes, perceptions, and practices of beneficiaries regarding micronutrient powder use. Emphasis was placed on identification of key messages and media for promotion of micronutrient powder in the future. The research team collaborated with two local artists in the refugee camp to draft revised micronutrient powder packaging and media based on socio-cultural information collected from in-depth interviews with beneficiaries. Feedback on those materials was then sought during the focus group discussions. Data analysis

Analysis of the data was an iterative, ongoing process. All interviews and focus group discussions were digitally recorded in local languages, translated into English, and transcribed. The analysis of the transcribed interviews and focus group discussions was done soon after transcriptions became available with the use of a codebook created by the two primary investigators. Codes were inductively generated within each of the community data sets using a “grounded” approach [12] and emerged from the participants’ descriptions of their perceptions, experiences, and recommendations with regard to the micronutrient powder program. The codes were then combined into broader categories or themes across the six ethnic communities to make them most pertinent for transferability of findings. Comparisons, trends, and paradoxes were then identified using data matrices as suggested by Maxwell [8], using Atlas.ti version 6.1 qualitative data computer software [13]. We sought to ensure transferability of the study findings by presentation of detailed, contextually rich results.

296

S. Kodish et al.

Stakeholder factors

Beneficiary factors Perceptions of MixMe™ Contraceptive

Superficial formative research

Interagency discordance

Weak CHW and scooper training

Insufficient social marketing

Lack of information provided

Experiences with MixMe ™ Product packaging

Threats to take MixMe™ at food distribution

Lack of information provided

Unknown ingredients Medicine rather than found

Low value of MixMe™

Consequences Distrust and confusion

Low uptake and usage

Causes GI troubles Leads to increased appetite

Cohort study

Implementing partner Implementing agency

Positive health effects

FIG. 1. Factors of low micronutrient powder uptake among beneficiaries. Arrow thickness denotes salience

Results Stakeholder and implementing partner views of MixMe

The low uptake of micronutrient powder resulted from numerous factors (fig. 1). Multiple implementing partners discussed feelings of unease with regard to the formative research that shaped this initiative. Conducted in March 2008, the acceptability study, which led to implementation of the micronutrient powder initiative, was felt to be incomplete by project staff and implementing partners. For example, the logo and packaging that were piloted in the study were different from those used in the actual micronutrient powder project. Also, the camp was approximately 75% Sudanese during the time of the acceptability study but had a Somali majority in February of the following year when the commodity was finally rolled out. The implementing partners also discussed challenges at an early stage in relation to the lack of partner cooperation during the project. Although our findings suggest that there was no overt competition among implementing partners, program staff did not perceive full unwavering cooperation in the implementation of the micronutrient powder project, which led to challenges during

commodity rollout. Multiple members of the program staff expressed frustration and confusion with partners’ lack of full support throughout the project. Disagreements also stemmed from decisions related to responsibility. In particular, the decision to have WFP, perceived by some other partners and beneficiaries as “the logistics agency of the camp,” deliver the commodity to the community, as opposed to the usual health providers in the camp (UNHCR* or IRC), caused feelings of unease among both implementing partners and beneficiaries. Without feelings of unity at the stakeholder level and among implementing partners, the project proceeded with its social marketing campaign. Using the findings from the limited acceptability inquiry, PSI commenced social marketing efforts that, due to issues largely out of its control (e.g., delays in disbursement of appropriated funds, rescheduled product launch date), lacked the coverage and intensity required to provide culturally appropriate information about the new commodity to the various subpopulations it was * Although UNHCR is by definition the primary agency in charge of camp coordination, data indicate that it is perceived by much of the refugee population as both the coordinating agency and one of the primary healthcare providers.

297

Qualitative inquiry of micronutrient powder program

meant to serve. Also, the communication plan soon ran into logistical delays related to ethical clearance of the cohort study by the Kenya Medical Research Institute and Johns Hopkins University School of Medicine, pushing back the actual product launch. Partners from various agencies talked about the delay of the launch as a major factor in low uptake of the commodity. If you were to plan a commercial launch of a product, what would be your dos and don’ts? And one thing that you never, ever want to do is to make a lot of noise and then not have the product available. And that is something that happened. Program staff, in-depth interview Once the product became available to the beneficiaries, questions arose from the beneficiaries about its qualities and utilities. Although the community health representatives and the “scoopers” (refugees who were responsible for distributing the food rations, including micronutrient powder, at food distribution centers) attempted to answer such questions, interview data suggest that the beneficiaries were oftentimes given incorrect or discouraging answers to micronutrient powder-related questions from these individuals. Although there were formal training sessions for some of the core community health representatives, particularly those who were involved in the cohort study assessments, a number of health workers were not trained in the same formal manner and felt unprepared to deal with community members’ questions. Lack of unanimous support from stakeholders, superficial formative research, complications of the cohort study, insufficient social marketing, and weak training of scoopers and community health representatives made the intervention a challenge from the outset. Refugee community members’ views of MixMe

Our investigation revealed numerous perceptions of MixMe in the community. Although some differences in opinions and understanding among communities did exist, the findings are synthesized in this section, because prominent, overarching themes emerged that were shared among all six communities, with little intersubgroup variation (fig. 1).

beneficiaries emphasized that they needed more education and awareness about the product itself: Usually when something is new that’s being introduced there is awareness being conducted, for example, explaining where this product came from, the agency concerns that brought this product. They will tell us what we will gain from this product, why is it beneficial, meaning what you will gain once you take the product, for example, anything that you eat has a role to play in the body or function you know. If you eat this product, what is it able to do in your body? Does it make you fat? Does it make you energetic? Or is it good for growth? I didn’t understand. I totally don’t know how this was brought to us because it is new and no one told us about it. I don’t know actually what this product is made from: wood, stone, what exactly are the ingredients of this product? Is it mixed with poison or medicines? Has it been expired or unmarketable to others so they give it to refugees to see if this product works? Seriously I want to know all of this because I don’t understand or know. We were never taught. Somali man, beneficiary, in-depth interview Without appropriate communications, the beneficiaries relied on information about MixMe from whatever sources they had available and came to their own conclusions about the product. They discussed their reliance on three things: the packaging of MixMe, their personal experiences using it, and their observations at distribution. Reliance on the packaging of MixMe for information

By relying on the packaging of MixMe, many beneficiaries began to understand the product very differently from its intended purpose. The beneficiaries thought that MixMe might be a contraceptive; they believed it might have been derived from deleterious ingredients; they also believed it could be a medicine. This category of issues emerged as the most salient theme related to low uptake of micronutrient powder among the beneficiary factors.

Lack of information on MixMe

As a result of delays that postponed the launch of the new product, insufficient social marketing efforts were carried out, which were not in parallel with the product launch. Also, because the population was constantly in flux, many refugees who received micronutrient powder during distribution missed the initial sensitization efforts made in fall 2008. As a result, the

FIG. 2. Image of family of three on MixMe box

298

MixMe as a contraceptive. The most prevalent beneficiary view of MixMe was that the product was for family planning. This idea stemmed in part from the MixMe packaging and logo and was directly related to core cultural values that communities held. Each community expressed distaste for the MixMe box, which had an image of a man, woman, and small child (fig. 2). To many, it indicated the promotion of small family size. A Sudanese Dinka mother of five discussed how the image of the family did not agree with the cultural norms of the Dinka community: That drawing should be changed because it’s against our norms; we believe and feel comfortable with huge family sizes, but that picture shows a family of three people, which is a contradiction to our tradition, hence the notion of a contraceptive. Sudanese Dinka woman, beneficiary, in-depth interview Just as the image of the family left an impression of family planning, so too did the color, size, and shape of the aluminum foil sachet (fig. 3), which resembled a condom package. A community leader who had been in the Kakuma Refugee Camp for 13 years in the Somali community explained the feelings of shame about the sachets because they are easily misunderstood to be condoms when in someone’s hand or pocket: Also the package looks like something put on for prevention, I mean condoms. I heard some people talking about it and they say they feel embarrassed when they have it in their pocket. Thinking that if someone saw them with it, they might think it was a condom. Somali man, community leader, in-depth interview

S. Kodish et al.

The use of contraception is either forbidden or heavily stigmatized among some Kakuma Refugee Camp communities, in accordance with religious values. Ingredients of MixMe.. Individuals from all communities stated their distaste for the perceived symbolism behind the main logo of MixMe (fig. 3). The respondents discussed concerns with the cartoon logo, especially in relation to its possible ingredients: The picture is showing that it has vitamins, but you see this genie picture? Even I don’t like talking to you about this picture. If they can change this picture, then it is OK. This one is not in line with our religion (Islam), you know, and people in our community were discouraged by this picture . . . some say that the MixMe is chopped bones of this picture. Somali Bantu man, community leader, in-depth interview The idea that the ingredients of the product might have been derived from the character in the cartoon logo emerged as a common theme. To many respondents, there was much ambiguity surrounding the ingredient base of the product. After experimentation due to a lack of trusted information, many individuals found that MixMe could be an adhesive when mixed with water, and the beneficiaries reported using it to patch roofs of dwellings and fix torn footwear. This finding further exacerbated a lack of coherent explanation surrounding the new product. This ambiguity of purpose proved to be a real concern, because food consumption in Muslim culture is guided by fundamental principles that are strictly adhered to, notably the concepts of halal and haram [14]. A leader of the Somali community summarized how a lack of information related to the ingredients of micronutrient powder can be a deterrent to those of Muslim faith: Yes, we are Muslim and according to our religion and culture for everything we use we must confirm the background of it whether it’s halal or haram. If it’s halal we use it very much and if it’s haram we dump it. Generally, things that grow on land are all halal. We can eat it, but in the case of things that are added or mixed together, like the vitamins and minerals, they need further explanation to make them (the ingredients) clearly explained to people. For example, people believe MixMe to be made of meat. Somali man, community leader, in-depth interview

FIG. 3. Cartoon logo on MixMe box and individual sachets

The lack of information on the ingredient base of MixMe was a catalyst for the search for knowledge by individuals and led to ambiguity regarding its composition and purpose for being introduced at the Kakuma Refugee Camp. MixMe as a medicine. Without a clear understanding

299

Qualitative inquiry of micronutrient powder program

of the nature and composition of the micronutrient powder, community members speculated whether it was a food or a medicine, which many people were averse to using in the absence of explicit illness: How do you think I can eat something which I don’t know about? . . . I mean just tell people the truth about whether it is a medicine or a food because MixMe is for anemia, I heard . . . so it is better if you just give it to people who have anemia. Ethiopian man, beneficiary, in-depth interview This perception was compounded by the flavorless nature of MixMe. A food product without flavor and in powder form was perceived to be medicinal by many community members. Reliance on personal experience with MixMe usage

Gaining understanding of the product was done not only through visual interpretation of the product, but also through experiential consumption. As a result of trials with the product itself, community members universally complained of gastrointestinal troubles (diarrhea, vomiting, nausea) and increased hunger, two deterrents to sustained use. MixMe was felt to promote hunger and appetite. Community members talked at length about the challenges they faced as a result of having increased appetites, in a setting in which inadequacy of food rations is often discussed as a foremost challenge: Most of the people were not sleeping well when taking MixMe and also they discovered in it a feeling of being hungry and that’s why they have decided not to use it regularly . . . because they can’t afford to buy extra food because if a person is of family size one, then he or she may use the ration up in 3 days . . . and this made them stop using MixMe on a regular basis. Somali woman, community leader, in-depth interview However, the beneficiaries also experienced positive health effects. The beneficiaries reported that micronutrient powder “reduced aches and pains,” “improved sleep,” and “eased pain during childbirth,” just to name a few of the cited positive effects. These perceptions were interspersed throughout the communities and added to the mixed messages to which people were exposed, contributing to overall confusion about the new commodity. Reliance on distribution for information

At food distribution, the scoopers handed out MixMe boxes and entertained questions that the beneficiaries had about the new product. Various accounts indicate

that the scoopers were forcing individuals to take MixMe through verbal threats, a counterproductive strategy that turned many off from the product. Many people disposed of the micronutrient powder upon exiting from the food distribution center. This created a situation in which many beneficiaries questioned the perceived value of the product, as others were discarding it in great quantities: When the uninterested beneficiaries receive a packet of MixMe, they dump it in the rubbish or on the streets so the interested people feel discouraged whenever they see this product being dumped outside and everyone is stepping on it. Somali woman, beneficiary, in-depth interview Freelist data indicate that MixMe was not a salient preventive measure or treatment for anemia. From the six different ethnic groups interviewed, only 11 of 72 individuals (15.3%) mentioned MixMe when asked to list all of the possible preventive measures or treatments for anemia; in fact, a large number of different measures (31) were mentioned as effective to address anemia. The most salient item mentioned by members of these six communities for prevention or treatment of anemia was to eat green vegetables (51.4%). Lack of trust

Resentment and distrust due to a delayed product launch, coupled with the measurements of the cohort study, which included blood drawing and anthropometric measurements at the time of the product launch, were associated with lingering suspicion surrounding the product’s effectiveness and safety. A beneficiary discusses this experience: I personally believe that the project of MixMe is doing research on whether the product will work or not, so they are testing on the refugees because they think we are homeless and anything we are given we will eat. That is very offensive really and it is an abuse of our rights as refugees. That is my major objection to this product. Somali Bantu man, beneficiary, in-depth interview Because the product had not been introduced to the Turkana host community or to the population at the Dadaab Refugee Camp in eastern Kenya, many community members believed they were being experimented on. Also, as a result, the beneficiaries became increasingly suspicious of the safety and value of the product, a major impediment to uptake and usage: Some say this product was not given to local Turkanas. We know all items of the food basket that we

300

get and that WFP distributes to Turkana sometimes, but when it comes to MixMe they (the Turkana) were excluded. Many people questioned this, and while WFP was not open to responding why the Turkanas are spared the micronutrient powder, people started to suspect the product. Congolese woman, beneficiary, in-depth interview

Discussion To our knowledge, this is the first in-depth, qualitative research inquiry into an ongoing micronutrient powder program in a refugee setting. The investigation illustrated both the challenges faced in implementing a new micronutrient powder program in a refugee setting, and the utility of qualitative research for understanding them. In qualitative research, data credibility is enhanced through the use of emergent design, triangulation, search for negative cases, and member checking, among other approaches [15]. Each of these methods was used in this study. The multiphase, iterative approach allowed new themes to be addressed as they arose. Methodological and investigator triangulation were used to develop and support emergent themes [16]. In addition, we refined our hypotheses as the inquiry advanced and actively searched for negative or disconfirming cases of our emerging theory [16, 17]. We employed the practice of member checking by returning to chosen participants to review transcripts with them after translation and transcription in order to ensure credible data [18]. The study findings highlight the importance of careful program design and planning, with emphasis on ensuring involvement and commitment from all partners and stakeholders from the outset, investigating the role of cultural factors, timing for the provision of information, messaging, and logistics to make an intervention work. The MixMe project was beset by challenges related in part to inadequate and intermittent information provided to the beneficiaries, leading them to seek information about the new product themselves. A reliance on personal experiences with MixMe and issues with its packaging led to confusion and deleterious rumors, resulting in decreased uptake of micronutrient powder at distribution points. Importantly, the low uptake of micronutrient powder at the Kakuma Refugee Camp seems to be the exception rather than the rule among micronutrient powder programs that have been implemented. Numerous micronutrient powder interventions conducted in other countries and settings have had auspicious results, including a high acceptance of and compliance

S. Kodish et al.

with the product among the beneficiaries* [19, 20]. For example, a micronutrient powder project targeting children 6 to 59 months of age in a refugee camp in Nepal and one in Bangladesh implemented as an emergency response, which targeted both children aged 6 to 59 months and pregnant and lactating women, saw more favorable acceptance than that observed at Kakuma, where the micronutrient powder was targeted to the general camp population [21, 22]. Combining micronutrient powder distribution with general distribution of food to be collected by a very diverse camp population made the situation at Kakuma very unique in terms of micronutrient powder programming. Despite previous successes with targeted micronutrient powder programs, our results indicate that high acceptability should not be assumed. De Pee and colleagues [19] suggest three key guidelines to follow to provide the best chance for acceptance: first, the packaging should be culturally appropriate and clear and self-explanatory with regard to content, target group, and methods and frequency of use; second, clear information, education, and communication (IEC) materials should be provided together with a good social marketing campaign, combined with adequate training given to those who will provide the product; and third, the mother or caretaker should receive the product with explicit instructions on its proper use in a relatively orderly distribution situation, to increase the chance of acceptance. Our results suggest that these three guidelines were not fully adhered to at the Kakuma Refugee Camp. Importance of substantial formative research

To effectively design and tailor a culturally appropriate public health intervention, thorough formative research is needed [23]. High-quality formative research is the basis for developing effective strategies, including communication channels, for behavior change and assists project staff in identifying, prioritizing, accessing, and targeting [24]. This research should incorporate an in-depth qualitative component and complementary quantitative methods. Our data suggest that the current project was hampered by an inadequate formative phase and the unanticipated challenges associated with preparing for an intervention with a diverse, transient population. A systematic approach to * Helen Keller International. Annual Nutrition Survey— Rohingya Refugee Cox’s Bazar, Bangladesh. United Nations High Commissioner for Refugees. Unpublished report, 2009; World Food Programme. Cross-sectional assessment at the end of the micronutrient powder intervention in the Cyclone Sidr affected area. Unpublished report, 2009; World Vision Mongolia. Effectiveness of home-based fortification of complementary foods with Sprinkles in an integrated nutrition program to address rickets and anemia. Unpublished report, 2005.

301

Qualitative inquiry of micronutrient powder program

formative research incorporating in-depth interviews, focus groups, and participant observation can allow implementers of micronutrient powder programs to better understand target populations in their natural settings, formulate culturally appropriate communication and promotion strategies, and foresee and address barriers to implementation before, rather than as, they arise. A formative phase also can also be an appropriate stage to identify and consider any stakeholder and partner involvement in a setting, including existing infant and young child feeding programs that might already be in place. Improvements to product packaging

The many concerns related to packaging transcended ethnic group. Not only were there concerns about a lack of appropriate, clear information on the micronutrient powder box and sachet, but there were also concerns stemming from what the images suggested. During formative research, local artists should be engaged to help develop culturally appropriate packaging for micronutrient powder in the targeted communities. This type of partnership has proven fruitful in other public health interventions [25]. Local artists are familiar with the colors, images, and fonts that are preferred by individuals in the community in question and can draft materials based on specific community feedback. For example, more appropriate packaging might have depicted larger family sizes, which would have been culturally resonant with the Sudanese Dinka community in particular. Social marketing and training

A diverse population necessitates strategic, targeted, and thorough social marketing [26]. Audience segmentation is a key social marketing principle and can permit the targeting of beneficiary populations with specific, culturally appropriate information. A thorough marketing plan needs to coincide with the product launch so as not to create expectations that go unmet. PSI, the large social marketing firm contracted for this project, indicated frustration that the product rollout occurred 5 months after demand creation was initiated. Selection of communication channels is essential. The beneficiaries discussed the importance of engaging community leaders and having their buy-in to micronutrient powder. Because social networks at the Kakuma Refugee Camp are so dense, social norms may be stronger and harder to change [27], highlighting how important the help from community leadership is when trying to attain system-level critical mass (i.e., micronutrient powder usage becomes self-sustaining) using a diffusion of innovations approach [28]. Many beneficiaries emphasized that they surely would have

used micronutrient powder had their local leaders done so. Targeting community leaders for early adoption should be considered in future interventions. Just as local leaders are a driving force behind community action at the Kakuma Refugee Camp, so too are community health workers. We found the need for more comprehensive training of community health workers and scoopers at the camp, not least because they were the front line of information dissemination in the MixMe project. If both community leaders and health workers do not buy in to a new health- and food-related project, then we cannot expect the beneficiaries to do so. The marketing firm further discussed problems related to a lack of contingency plan to cater to emerging fiscal needs associated with the delay of the launch. Cost considerations are important, as they limit the extent to which a marketing campaign can be carried out. The firm explained how the untimely release of funds seriously impeded social marketing efforts at the Kakuma Refugee Camp. Such issues should be addressed early and agreed upon by all parties involved. Other lessons learned

Any successful public health intervention requires careful planning from the start. It is clear from the data in this inquiry that disparate opinions and a lack of agreement among stakeholders were impediments to uptake. In the planning stages of an intervention, particularly in a protracted situation such as that of the Kakuma Refugee Camp, time is available to develop and implement strategies that ensure clear agreement and full cooperation among stakeholders. This case study highlights the importance of process evaluation in an intervention trial. During process evaluation, documentation and analysis of early program development and implementation should be carried out to determine where expected output was actually produced [29]. More closely monitoring the fidelity and quality of implementation, in addition to the acceptability and actual usage of micronutrient powder, would have been beneficial to this project’s success. Monitoring actual usage rather than just uptake of micronutrient powder at distribution points should be a goal of future interventions so as to better link behavior with health outcomes. A summary of recommendations is outlined in table 3. Limitations

The results of this inquiry are specific to the Kakuma Refugee Camp, and transferability of findings [18] must be applied with caution. However, issues that emerged from this inquiry should be taken into account in other settings where implementation of micronutrient powder programs is being considered. Another

302

S. Kodish et al.

TABLE 3. Summary of recommendations Next steps at Kakuma Refugee Camp » The results of this study should be presented to agency stakeholders that serve Kakuma for discussion and feedback prior to implementing a new micronutrient powder project. » If a new project is implemented, it should be targeted (i.e., to children 24 to 59 months of age and schoolchildren), rather than a general program. » If a new micronutrient powder product were to be introduced, it would have to be remarketed with completely new packaging, logo, and flavor modifications in line with cultural themes valued by community members. » A new micronutrient powder product should be supported by an extensive social marketing campaign that promotes the product through locally designed, culturally appropriate materials by refugee artists. » A new micronutrient powder product should be promoted and used by community leaders and important health staff (following a diffusion of innovations model). » Extensive training regarding micronutrient powder needs to be provided to the community health workers prior to its introduction. » The potential for introduction of a new micronutrient powder to local markets at a nominal cost should be considered. General recommendations for micronutrient powder programming » Both in initial program design and during implementation, stakeholder involvement and collaboration are pivotal. The responsibilities of each stakeholder and partner involved need to be clearly defined and agreed upon. Resolution of discordant stakeholder views should be made a priority throughout programming. » At the front end of the project, detailed, multistage formative research that spans multiple months is crucial. Culturally appropriate communication and promotion strategies and product packaging need to be developed and pretested during this phase. » Expansive social marketing using culturally appropriate materials should be an essential component. » Project staff and community health workers require proper training prior to and throughout the program, and their performance needs to be carefully monitored. » Once a micronutrient powder program is under way, extensive process evaluation should be implemented to assess program implementation fidelity, in addition to reach and dose. » Research is warranted to establish and document the causal chain of micronutrient powder from uptake and distribution, to utilization in the family food supply, to intake at the individual level. The development of simple, valid, and reliable indicators of micronutrient powder compliance at the individual level in a program condition may be a prerequisite. » Impact studies held in conjunction with micronutrient powder programming need to be carefully planned and unintended consequences (e.g., suspicions that might arise from the taking of anthropometric measurements and blood drawing) considered.

consideration is the nonrandom sample of participants in this study. Following established qualitative research principles, each participant in this sample was chosen for his or her contribution to the emerging theory to ensure comprehensive, complete, and saturated accounts of the MixMe project [30].

Conclusions This study serves as an important case study for other programs wishing to implement successful micronutrient powder strategies in refugee settings and elsewhere. Although the generalizability of the present findings may be limited due to the unique context of the Kakuma Refugee Camp, general lessons learned, including the need for improved formative research in the design phase, early coordination among stakeholders, and careful process monitoring, can be applied to most programs. In addition, further promotion of the use of qualitative approaches in various phases of a micronutrient powder program, such as formative

research, process monitoring, and impact evaluation, is warranted.

Acknowledgments The present research was supported by DSM as part of a joint initiative by the WFP-DSM partnership known as “Improving Nutrition—Improving Lives.” The authors wish to acknowledge Joris van Hees, WFP Rome, Christine Klotz, WFP Kenya, and UNHCR for reviewing the draft and providing thoughtful comments. We wish to thank the WFP Kakuma staff for their tremendous support during the fieldwork portion of this study, as well as the hard-working data collection team: Arbab Khims Adam, Debela Gedefa, Elizabeth Akon, Faysal Ahmed, Hussein Adan Hassan, Miyom Manas Wal, Mohamud Muridi Saidi, Nul Deng Lual, Rubera Bisengimana Evarest, Stephen Abboud Arial, and Temesgen Feleke. A Harry D. Kruse Publication Award in Human Nutrition is gratefully acknowledged.

Qualitative inquiry of micronutrient powder program

303

References 1. Berry-Koch A, Moench R, Hakewill P, Dualeh M. Alleviation of nutritional deficiency diseases in refugees. Food Nutr Bull 1990;12:106–12. 2. Dye TD. Contemporary prevalence and prevention of micronutrient deficiencies in refugee settings worldwide. J Refug Stud 2007;20:108–19. 3. de Pee S, Kraemer K, Van Den Briel T, Boy E, Grasset C, Moench-Pfanner R, Zlotkin S, Bloem MW. Quality criteria for micronutrient powder products: Report of a meeting organized by the World Food Programme and Sprinkles global health initiative. Food Nutr Bull 2008;29:232–41. 4. Christofides A, Schauer C, Zlotkin S. Iron deficiency anemia among children: Addressing a global public health problem within a Canadian context. Pediatr Child Health 2005;10:597–601. 5. Dewey KG, Yang Z, Boy E. Systematic review of metaanalysis of home fortification of complementary foods. Matern Child Nutr 2009;5:283–321. 6. Hyder S, Zlotkin SH, Haseen F, Zeng L. Efficacy of daily vs. weekly home fortification of weaning foods with Sprinkles among infants and young children in Dhaka, Bangladesh. In: Workshop Proceedings from the National Workshop on Home Fortification of Weaning Food with Sprinkles: A new strategy to control anaemia among infants and young children. Dhaka: Bangladesh Rural Advancement Committee, 2004:1–14. 7. Lundeen E, Schueth T, Toktobaev N, Zlotkin S, Hyder SM, Houser R. Daily use of Sprinkles micronutrient powder for 2 months reduces anemia among children 6 to 36 months of age in the Kyrgyz Republic: A clusterrandomized trial. Food Nutr Bull 2010;31:446–60. 8. Maxwell J. Qualitative research design: An iterative approach, 2nd ed. Thousand Oaks, Calif, USA: Sage Publications, 2005. 9. Mayes N, Pope C. Assessing quality in qualitative research. BMJ 2000;320:50–2. 10. Weller SC, Romney AK. Structured interviewing. Newbury Park, Calif, USA: Sage Publications, 1988. 11. Kim YM, Heerey M, Kols A. Factors that enable nurse-patient communication in a family planning context: A positive deviance study. Int J Nurs Stud 2008; 45:1411–21. 12. Glaser B, Strauss A. Awareness of dying. Chicago, Ill, USA: Aldine, 1965. 13. Atlas.ti (Version 6.1) [Computer Software]. Berlin: Scientific Software Development. 14. Kocturk T. Food rules in the Koran. Scand J Food Nutr 2002;46:137–40. 15. Creswell JW. Qualitative inquiry and research design: Choosing among five approaches, 2nd ed. Thousand Oaks, Calif, USA: Sage Publications, 2007. 16. Miles MB, Huberman AM. Qualitative data analysis, 2nd ed. Thousand Oaks, Calif, USA: Sage Publications, 1994.

17. Glesne C, Peskin A. Becoming qualitative researchers: An introduction. White Plains, NY, USA: Longman, 1992. 18. Lincoln AS, Guba EG. Naturalistic inquiry. Beverly Hills, Calif, USA: Sage Publications, 1985. 19. de Pee S, Moench-Pfanner R, Martini E, Zlotkin SH, Darnton-Hill I, Bloem MW. Home fortification in emergency response and transition programming: Experiences in Aceh and Nias, Indonesia. Food Nutr Bull 2007;28:189–97. 20. Menon P, Ruel MT, Loechl CU, Arimond M, Habicht J, Pelto G, Michaud L. Micronutrient Sprinkles reduce anemia among 9- to 24-mo-old children when delivered through an integrated health and nutrition program in rural Haiti. J Nutr 2007;137:1023–30. 21. Bilukha O, Howard C, Wilkinson C, Bamrah S, Husain F. Effects of multimicronutrient home fortification on anemia and growth in Bhutanese refugee children. Food Nutr Bull 2011;32:263–75. 22. Rah JH, de Pee S, Halati S, Parveen M, Mehjabeen SS, Steiger G, Bloem MW, Kraemer K. Provision of micronutrient powder in response to the Cyclone Sidr emergency in Bangladesh: Cross-sectional assessment at the end of the intervention. Food Nutr Bull 2011;32:276–84. 23. Strolla LO, Gans KM, Risica PM. Using qualitative and quantitative formative research to develop tailored nutrition intervention materials for a diverse lowincome audience. Health Educ Res 2006;21:465–76. 24. Higgins DL, O’Reilley K, Tashima N, Crain C, Beeker C, Goldbaum G, Sterk Elifson C, Galavotti C, GuentherGrey C. Using formative research to lay the foundation for community level HIV prevention efforts: An example from the AIDS community demonstration projects. Public Health Rep 1995;111:29–35. 25. Panda S, Jana M, Sourav Das R, Pahari S, Maruf S, Das R, Varma R, Mandal A, Gere D. Art and public health: Report on development of a unique community-based approach to reduce HIV/AIDS stigma in villages of West Bengal. 1998. Available at: http://makeartstopaids.org/ publications.htm. Accessed 13 May 2011. 26. Manoff RK. Social marketing: New imperative for public health. New York: Praeger, 1985. 27. Berkman LF. Assessing the physical health effects of social networks and social support. Public Health 1984;5:413–32. 28. Rogers EM. Diffusion of innovations. Glencoe, Ill, USA: Free Press, 1962. 29. Mukoma W, Flisher A, Ahmed N, Jansen S, Mathews C, Klepp K, Schaalma H. Process evaluation of a schoolbased HIV/AIDS intervention in South Africa. Scand J Public Health 2009;39(suppl 2):37–47. 30. Morse JM. Qualitative generalizability. Qual Health Res 1999;9:5–6.