Comparing estimated cost-effectiveness of micronutrient intervention programs using primary and secondary data: evidence from Cameroon.

Designing a cost-effective portfolio of micronutrient intervention programs is complex and generally undertaken with limited data. We developed the MINIMOD-Secondary Data (MINIMOD-SD) tool, which uses household consumption and expenditure survey data and other secondary data to estimate apparent nutrient intakes and model the effectiveness and cost-effectiveness of micronutrient intervention programs. We present the SD tool methodology and results in the context of Cameroon, with a particular focus on vitamin A (VA) for children and folate for women of reproductive age (WRA). We compared the MINIMOD-SD tool estimates with those of the full MINIMOD tool, which uses 24-h dietary recall data. The SD tool consistently underestimated folate intake among women (median (IQR): 230 (143,352) versus 303 (244,367) µg dietary folate equivalents (DFEs)/day) and especially VA among children (141 (64,279) versus 227 (102,369)). Qualitatively, however, the two tools were generally consistent in predicted subnational patterns of micronutrient adequacy and identification of effective and cost-effective (cost per child/WRA moving from inadequate to adequate intake) interventions. Secondary data and the MINIMOD-SD tool can provide policymakers with information to qualitatively assess deficiency risks and identify cost-effective interventions. However, accurately quantifying individual-level deficiency or dietary inadequacy and intervention effectiveness and cost-effectiveness will likely require individual-level dietary data and biomarker measurements.

Factors associated with the performance of routine health information system in Yaoundé-Cameroon: a cross-sectional survey.

BACKGROUND: Routine Health Information Systems (RHIS) of low-income countries function below the globally expected standard, characterised by the production and use of poor-quality data, or the non-use of good quality data for informed decision making. This has negatively influenced the health service delivery and uptake. This study focuses on identifying the factors associated with the performance of RHIS of the health facilities (HF) in Yaoundé, so as to guide targeted RHIS strengthening. METHODS: A HF-based cross-sectional study in the 6 health districts (HDs) of Yaoundé was conducted. HFs were chosen using stratified sampling with probability proportional to size per HD. Data were collected, entered into Microsoft Excel 2013 and analysed with IBM- SPSS version 25. Consistency of the questionnaire was measured using Cronbach's alpha coefficient. Pearson's chi-square (and Fisher exact where relevant) tests were used to establish relationships between qualitative variables. Associations were further quantified using unadjusted Odd ratio (OR) for univariable analysis and adjusted odds ratio (aOR) for multivariable analysis with 95% confidence interval (CI). A p-value of less than 0.05 was considered statistically significant. RESULTS: Of 111 selected HFs; 16 (14.4%) were public and 95 (85.6%) private. Respondents aged 24-60 years with an average of 38.3 ± 9.3 years; 58 (52.3%) males and 53(47.7%) females. Cronbach's alpha was 0.96 (95%CI: 0.95-0.98, p < 0.001), proving that the questionnaire was reliable in measuring RHIS performances. At univariable level, the following factors were positively associated with good performances: supportive supervision (OR = 3.03 (1.1, 8.3); p = 0.02), receiving feedback from hierarchy (OR = 3.6 (0.99, 13.2); p = 0.05), having received training on health information (OR = 5.0 (1.6, 16.0); p = 0.003), and presence of a performance evaluation plan (OR = 3.3 (1.4, 8.2), p = 0.007). At multivariable level, the only significantly associated factor was having received training on health information (aOR = 3.3 (1.01, 11.1), p = 0.04). CONCLUSION: Training of health staff in the RHIS favors RHIS good performance. Hence, emphasis should be laid on training and empowering staff, frequent and regular RHIS supervision, and frequent and regular feedback, for an efficient RHIS strengthening in Yaoundé.

A qualitative study to understand how Ebola Virus Disease affected nutrition in Sierra Leone-A food value-chain framework for improving future response strategies.

BACKGROUND: This study sought understand how the 2014-2016 EVD Virus Disease (EVD) outbreak impacted the nutrition sector in Sierra Leone and use findings for improving nutrition responses during future outbreaks of this magnitude. METHODOLOGY: This qualitative study was iterative and emergent. In-depth interviews (n = 42) were conducted over two phases by purposively sampling both key informants (n = 21; government stakeholders, management staff from United Nations (UN) agencies and non-governmental organizations (NGO)), as well as community informants (n = 21; EVD survivors, health workers, community leaders) until data saturation. Multiple analysts collaborated in a team-based coding approach to identify key themes using Dedoose software. Findings are presented as both quotations and tables/figures. RESULTS: The EVD outbreak effects and the related response strategies, especially movement restriction policies including 21-day quarantines, contributed to disruptions across the food value-chain in Sierra Leone. System-wide impacts were similar to those typically seen in large-scale disasters such as earthquakes. Participants described an array of direct and indirect effects on agricultural production and food storage and processing, as well as on distribution, transport, trade, and retailing. Secondary data were triangulated by interviews which described the aggregate negative effect of this outbreak on key pillars of food security, infant and young child feeding practices, and nutrition. During the humanitarian response, nutrition-specific interventions, including food assistance, were highly accepted, although sharing was reported. Despite EVD impacts across the entire food value-chain, nutrition-sensitive interventions were not central to the initial response as EVD containment and survival took priority. Culturally-appropriate social and behavior change communications were a critical response component for improving health, nutrition, and hygiene-related behaviors through community engagement. CONCLUSIONS: Infectious diseases such as EVD have far-reaching effects that impact health and nutrition through interrelated pathways. In Sierra Leone, the entire food value-chain was broken to the extent that the system-wide damage was on par with that typically resulting from large natural disasters. A food value-chain approach, at minimum, offers a foundational framework from which to position nutrition preparedness and response efforts for outbreaks in similar resource constrained settings.

Monitoring of the National Oil and Wheat Flour Fortification Program in Cameroon Using a Program Impact Pathway Approach.

BACKGROUND: Since 2011 Cameroon has mandated the fortification of refined vegetable oil with vitamin A and wheat flour with iron, zinc, folic acid, and vitamin B-12. In 2012, measured fortification levels for flour, and particularly oil, were below target. OBJECTIVES: We assessed Cameroon's food fortification program using a program impact pathway (PIP) to identify barriers to optimal performance. METHODS: We developed a PIP through literature review and key informant interviews. We conducted interviews at domestic factories for refined vegetable oil (n = 9) and wheat flour (n = 10). In 12 sentinel sites distributed nationally, we assessed availability and storage conditions of fortified foods in markets and frequency of consumption of fortified foods among women and children (n = 613 households). Food samples were collected from factories, markets, and households for measurement of micronutrient content. RESULTS: Two-thirds of factories presented quality certificates for recent premix purchases. All factories had in-house capacity for micronutrient analysis, but most used qualitative methods. Industries cited premix import taxes and access to external laboratories as constraints. Mean vitamin A levels were 141% (95% CI: 116%, 167%), 75% (95% CI: 62%, 89%), and 75% (95% CI: 60%, 90%) of target in individual samples from factories, markets, and households, respectively. Most industry flour samples appeared to be fortified, but micronutrient levels were low. Among composite flour samples from markets and households, the mean iron and zinc content was 25 mg/kg and 43 mg/kg, respectively, ∼45% of target levels; folic acid (36%) and vitamin B-12 (29%) levels were also low. In the previous week, the majority of respondents had consumed "fortifiable" oil (63% women and 52% children) and wheat flour (82% women and 86% children). CONCLUSIONS: In Cameroon, oil fortification program performance appears to have improved since 2012, but fortification levels remain below target, particularly for wheat flour. Consistent regulatory monitoring and program support, possibly through premix procurement and micronutrient analysis, are needed.

Consensus building around nutrition lessons from the 2014-16 Ebola virus disease outbreak in Guinea and Sierra Leone.

There are important lessons learned from the 2014-16 Ebola virus disease outbreak in West Africa. However, there has not been a systematic documentation of nutrition lessons specifically. Therefore, this study sought to generate multiple stakeholder perspectives for understanding the nutrition challenges faced during the Ebola virus disease outbreak, as well as for consensus building around improved response strategies. Participatory workshops with 17 and 19 participants in Guinea and Sierra Leone, respectively, were conducted in February 2017. Workshops followed the Nominal Group Technique, which is a methodological approach for idea generation and consensus building among diverse participants. Those findings were triangulated with qualitative interview data from participants representing government, United Nations bodies, civil society, non-governmental organizations and local communities in both Guinea (n = 27) and Sierra Leone (n = 42). (1) Reduced health system access and utilization, Poor caretaking and infant and young child feeding practices, Implementation challenges during nutrition response, Household food insecurity and Changing breastfeeding practices were five nutrition challenges identified in both Guinea and Sierra Leone. (2) Between settings, 14 distinct and 11 shared organizational factors emerged as facilitators to this response. In Sierra Leone, participants identified the Use of Standard Operating Procedures and Psychosocial counselling, whereas in Guinea, Hygiene assistance was distinctly important. Political will, Increased funding, Food assistance and to a lesser extent, Enhanced coordination, were deemed 'most important' response factors. (3) The top nutrition lessons learned were diverse, reflecting those of nutrition policy, programme implementation, community activity and household behaviours. Disease outbreaks pose widespread nutrition challenges to populations in resource-constrained settings where global health security is not a guarantee. These findings should be considered for emergency nutrition preparedness and inform evidence-based priority setting in the post-Ebola virus context of Guinea and Sierra Leone.

Correction: Implications of the Ebola virus disease outbreak in Guinea: Qualitative findings to inform future health and nutrition-related responses.

[This corrects the article DOI: 10.1371/journal.pone.0202468.].

Implications of the Ebola virus disease outbreak in Guinea: Qualitative findings to inform future health and nutrition-related responses.

INTRODUCTION: Due to the close relationship between EVD and nutrition, the humanitarian community implemented various nutrition-specific and -sensitive interventions to stem the Ebola Virus Disease (EVD) outbreak in West Africa. Little, however, is known about stakeholder and community members' perspectives toward this response in Guinea. Therefore, we aimed to firstly understand how EVD may have influenced the nutrition situation; and secondly to assess the perceived acceptability and effectiveness of the nutrition response. MATERIALS AND METHODS: Using 27 in-depth interviews conducted in April-May 2016, this descriptive, qualitative study had three iterative phases in an emergent design. Phase 1 explored the perceptions of 11 high-level policy and management staff. Phase 2 assessed the views of 16 community members, survivors, and front-line workers. Phase 3 compared the qualitative findings to relevant nutrition indicators from secondary data for final interpretations. A systematic, team-based coding approach using Dedoose software identified key themes during textual analysis. RESULTS: Overall, several plausible pathways through an interrelated network of bio-social factors help describe EVD impacts on the nutrition situation of Guinea. At a basic level, complex social dimensions of health, response unpreparedness, and market disruptions were perceived to be major determinants affecting the nutrition situation, especially among IYC. At an underlying level, household food security was negatively impacted, along with weakened care-seeking practices, IYC feeding practices, and coping strategies. Consequently, treatment coverage for childhood illnesses and IYC diets were negatively impacted during the outbreak. In hindsight, most participants had positive perceptions toward the overall EVD response, but described salient considerations for improving upon this nutrition response during future outbreaks. DISCUSSION: This study highlighted the complex web of inter-related factors through which EVD was perceived to impact the nutrition situation in Guinea. Considering the multi-level social and behavioral dimensions of health and nutrition is critical for effectively responding to infectious disease outbreaks.

Determinants of Stunting, Wasting, and Anemia in Guinean Preschool-Age Children: An Analysis of DHS Data From 1999, 2005, and 2012.

Wasting, stunting, and anemia are persistent and important forms of malnutrition in preschool-age children in the less developed world, in particular the Republic of Guinea, which was the site of a large outbreak of Ebola virus disease in 2014 to 2015. We analyzed data from 3 Demographic and Health Surveys done in Guinea in 1999, 2005, and 2012 to identify possible determinants of wasting, stunting, and anemia. All analyses, both bivariate and multivariate, were carried out separately for each of 3 age groups: less than 6 months, 6 to 23 months, and 24 to 59 months. Variables found statistically significantly associated with stunting, wasting, or anemia in bivariate analysis were placed in an age-specific logistic regression model for that outcome. Overall, anthropometric indices were available for 9228 children and hemoglobin concentrations were available for 5681 children. Logistic regression found relatively few variables associated with nutrition outcomes in children younger than 6 months. More variables were associated with nutrition outcomes in children aged 6 to 23 months. Such variables measured a wide variety of conditions, including estimated birth size, child health and nutritional status, child caring practices, mother's nutritional and health status, and household water source and sanitation. A similarly broad range of variables was statistically significantly associated with one or more nutrition outcomes in children aged 24 to 59 months. Few of the standard infant and young child feeding indicators were associated with any nutrition outcome. Improvement in the nutritional status of young children in Guinea may require a broad range of nutrition and health interventions.

Childhood anemia in Rural Haiti: the potential role of community health workers.

BACKGROUND: Childhood iron deficiency anemia (IDA) is an important contributor to under-five mortality in the developing world. There is evidence that Community Health Worker (CHW) delivered programs to increase maternal knowledge of child health practices may decrease childhood IDA. This study reports findings on the association between a long standing CHW intervention and childhood anemia status in rural Haiti. METHODS: Using structural equations and mediation analyses on data from a household-based survey of 621 mother/child dyads, we tested the hypothesis that CHW would have a direct positive effect on maternal knowledge and an indirect effect on childhood anemia in rural Haiti. RESULTS: CHW contact was significantly associated with maternal knowledge of key child health practices (ß = 0.193, SE = 0.058, p = 0.001). However, knowledge was not associated with childhood anemia (ß = -0.008, SE = 0.009, p = 0.382). Maternal knowledge categories significantly affected by CHW contact included diarrheal prevention knowledge (ß = 0.111, SE = 0.045, p = 0.013) and signs of malnutrition (ß = 0.217, SE = 0.071, p = 0.002). There was no significant association with knowledge of vitamin A and iron rich foods (ß = 0.057, SE = 0.032, p = 0.074), which is the intervention most likely to impact childhood anemia. In all path models tested, we identified the control variables low household socio-economic status, mothers' anemia status, and child's age less than 24 months as significant predictors of childhood anemia. CONCLUSIONS: CHWs delivered interventions are associated with improved maternal knowledge of child health practices in rural Haiti; however, this knowledge is not associated with improved childhood anemia. Concurrently with CHW-delivered programs, interventions household poverty are implied to impact childhood health outcomes in resource poor settings.

Promotion of zinc tablets with ORS through child health weeks improves caregiver knowledge, attitudes, and practice on treatment of diarrhoea in Nigeria.

We examined whether the Maternal, Newborn and Child Health Weeks (MNCHW) in Nigeria would present an opportunity to raise awareness of and demand for the use of zinc and ORS in the treatment for diarrhoea, guided by a conceptual framework designed to assess three theoretical underpinnings (characteristics and performance standard of the health workers, potential reach, and intensity of the intervention), along the impact pathway. Zinc and ORS with education for their appropriate use during the next diarrhoeal episode were delivered as part of the November 2010 and May 2011 MNCHW. On the day of but before participating in MNCHW activities, semi-structured interviews were used for collecting information on knowledge, attitudes, and practice (KAP) relating to diarrhoea from 602 caregivers with children aged less than five years. Forty-eight health workers were also interviewed. Nearly all health workers (98%) correctly mentioned the dosage of zinc while only 58% correctly stated the preparation of ORS. The proportion of caregivers with knowledge on the treatment for diarrhoea increased from 46.4% in November 2010 pre-MNCHW to 71.3% in May 2011 pre-MNCHW interviews (p<0.001). More caregivers correctly mentioned the dosage of zinc (80.9%) and stated the preparation of ORS (88.8%) in the November 2010 exit interview immediately after the MNCHW encounter compared to the levels a few months later in the home follow-up visit (53.1% and 37.4% respectively). After attending both rounds of November 2010 and May 2011 MNCHW, caregivers' knowledge on the treatment of diarrhoea increased seven times compared to the caregivers who attended the May 2011 MNCHW only (OR=7.0, p<0.001). Additionally, caregivers were 40% less likely to seek advice outside the home in the treatment for diarrhoea if they had attended both the MNCHWs than if they had attended the May 2011 MNCHW only (OR=0.6, p<0.029). We conclude that providing opportunities for caregivers to receive a sample of zinc and ORS and to learn about its use in the treatment for diarrhoea, from trained health workers during MNCHW, has the potential to increase KAP relating to the use of zinc and ORS in the treatment for diarrhoea and for future diarrhoeal episodes.

Determinants of institutional delivery among women of childbearing age in rural Haiti.

Institutional delivery is an important factor associated with reduced maternal mortality rate (MMR). MMR in Haiti is high (350 per 100,000) and institutional delivery is low-just over 25 % of women delivered at a health facility in 2010. There also exists substantial rural-urban disparity in delivery with more hospital deliveries in urban than in rural areas. We aimed to study the prevalence and determinants of institutional delivery in a sample of women of childbearing age in rural Haiti. The study took place in Fond des Blancs and Villa, as part of a baseline assessment undertaken prior to implementation of a maternal, child health, nutrition, and water and sanitation program. From October to November 2011, women 15-49 years old (N = 575) were selected using a cross-sectional two-stage sampling strategy. We used descriptive and multivariate logistic regression analyses to assess the prevalence of and factors associated with institutional delivery. The prevalence of institutional delivery was 45.4 %; a rate higher than the national average of 25 %. In adjusted analyses, correlates of institutional delivery were younger maternal age (25 years and younger) (OR 1.82; CI 1.15, 2.90; P = 0.0112), antenatal care receipt (OR 3.70; CI 1.84, 7.43; P = 0.0003) and those who were poor according to our poverty index score classification (OR 2.04; CI 1.13, 3.69; P = 0.0187). This study shows that increased hospital delivery is likely explained by accessibility to antenatal care. Programs that improve access to antenatal care, with concurrent efforts to address structural inequalities that drive socio-economic deprivation, are likely critical to increasing institutional delivery.

Précis of nutrition of children and women in Haiti: analyses of data from 1995 to 2012.

Between 1995 and 2012, many surveys including child and maternal nutrition indicators were conducted in Haiti. While many questions emerged from the results of those surveys, they have remained unanswered, in particular as they pertain to the determinants of poor children's and women's nutrition in Haiti. The purpose of this paper is to fill that gap and provide policymakers, program managers, and readers interested in nutrition issues in Haiti with information on (1) the trends and determinants of infant and young child feeding and food practices; (2) micronutrient deficiencies among children and women; (3) the status of severe acute malnutrition in children; (4) associations among women's empowerment, access to health care, water, and sanitation and child nutrition; (5) the current community-based early child care and nutrition initiatives; and (6) the status of nutrition governance in the country. By looking at many sources of data, including previously published and new data, we provide insight into major predictors of child malnutrition and associations among child feeding practices, maternal nutrition, and child growth outcomes. We also show that important progress has been made recently in the governance of nutrition programs and in child and maternal nutrition indicators, a result of effective evidence-based advocacy, partnerships, and design, implementation, and scale-up of nutrition-specific and sensitive interventions.

Acculturation and nutritional health of immigrants in Canada: a scoping review.

Although recent immigrants to Canada are healthier than Canadian born (i.e., the Healthy Immigrant Effect), they experience a deterioration in their health status which is partly due to transitions in dietary habits. Since pathways to these transitions are under-documented, this scoping review aims to identify knowledge gaps and research priorities related to immigrant nutritional health. A total of 49 articles were retrieved and reviewed using electronic databases and a stakeholder consultation was undertaken to consolidate findings. Overall, research tends to confirm the Healthy Immigrant Effect and suggests that significant knowledge gaps in nutritional health persist, thereby creating a barrier to the advancement of health promotion and the achievement of maximum health equity. Five research priorities were identified including (1) risks and benefits associated with traditional/ethnic foods; (2) access and outreach to immigrants; (3) mechanisms and coping strategies for food security; (4) mechanisms of food choice in immigrant families; and (5) health promotion strategies that work for immigrant populations.

Protecting and improving breastfeeding practices during a major emergency: lessons learnt from the baby tents in Haiti.

PROBLEM: The 2010 earthquake in Haiti displaced about 1.5 million people, many of them into camps for internally displaced persons. It was expected that disruption of breastfeeding practices would lead to increased infant morbidity, malnutrition and mortality. APPROACH: Haiti's health ministry and the United Nations Children's Fund, in collaboration with local and international nongovernmental organizations, established baby tents in the areas affected by the earthquake. The tents provided a safe place for mothers to breastfeed and for non-breastfed infants to receive ready-to-use infant formula. Such a large and coordinated baby tent response in an emergency context had never been mounted before anywhere in the world. LOCAL SETTING: Baby tents were set up in five cities but mainly in Port-au-Prince, where the majority of Haiti's 1555 camps for displaced persons had been established. RELEVANT CHANGES: Between February 2010 and June 2012, 193 baby tents were set up; 180 499 mother-infant pairs and 52 503 pregnant women were registered in the baby tent programme. Of infants younger than 6 months, 70% were reported to be exclusively breastfed and 10% of the "mixed feeders" moved to exclusive breastfeeding while enrolled. In 2010, 13.5% of registered infants could not be breastfed. These infants received ready-to-use infant formula. LESSONS LEARNT: Thanks to rapid programme scale-up, breastfeeding practices remained undisrupted. However, better evaluation methods and comprehensive guidance on the implementation and monitoring of baby tents are needed for future emergencies, along with a clear strategy for transitioning baby tent activities into facility and community programmes.

Prevalence and Risk Factors of Anemia among Children 6-59 Months Old in Haiti.

Anemia has serious consequences on child growth, development, and survival. This study was conducted in Fond des Blancs and Villa, Haiti, to assess the prevalence of childhood anemia and its risk factors in order to inform program design. Children 6-59 months old (n = 557) were selected using a cross-sectional multistage sampling methodology. Hemoglobin was measured using the HemoCue technique. Descriptive and multivariate analyses were performed to determine prevalence and factors associated with anemia. The prevalence of childhood anemia was 38.8% (23.9% mild, 14.7% moderate, and 0.2% severe). Mean hemoglobin was 11.2 ± 1.2 g/dL. Variables associated with child anemia were age less than 24 months (OR = 2.6; P = 0.000), stunting (OR = 2.2; P = 0.005), and mother's low hemoglobin level (OR = 1.8; P = 0.011). Anemia among young children in Fond des Blancs and Villa is a public health problem. Predictors of child anemia in this region include child's age, stunting, and mother's anemia. Interventions and strategies aimed at addressing effectively anemia in this population must therefore target mothers and children under two years of age.

Predictors of anemia among haitian children aged 6 to 59 months and women of childbearing age and their implications for programming.

BACKGROUND: The Haitian National Nutrition Policy prioritizes prevention and treatment of anemia among mothers and young children, but there are few available data to support planning for scale-up of anemia interventions. OBJECTIVE: To describe the prevalence and predictors of anemia among Haitian women (15 to 49 years) and children (6 to 59 months) and to draw implications for national nutrition programming. METHODS: Descriptive and univariate analyses and multivariate logistic regression models were performed using data from the nationally representative Haitian Demographic Health Survey 2005/06. RESULTS: The prevalence of mild (hemoglobin 11.0 to 11.9 g/dL), moderate (hemoglobin 8.0 to 10.9 g/dL), and severe (hemoglobin < 8.0 g/dL) anemia was 19.2%, 21.7%, and 4.4%, respectively, among women aged 15 to 49 years and 22.9%, 33.9%. and 2.2% among children aged 6 to 59 months. Unexpectedly anemia was more prevalent in urban women (54.4 %) and children (65.1%) than in rural women (43.1%, p < .001) and children (55.7%, p = .004). In multivariate regression models, factors associated with anemia among urban women (birth spacing, p = .027; overweight BMI, p < .001; education level, p = .022) were different from those in rural women (wealth quintile, p < .05; employment, p = .003). Anemia in urban and rural children aged 6 to 59 months increased with child age (p < .05) and maternal anemia status (p = .004; p < .001). Female sex (p = .007) and maternal overweight (p = .009) were associated with reduced risk of anemia in rural children only. CONCLUSIONS: Anemia among Haitian young children and women of childbearing age is a severe public health problem. The findings suggest the need for context-specific rural and urban strategies, reinforcement of anemia prevention in health services reaching women of childbearing age, and targeted interventions for young children.

Child malnutrition in Haiti: progress despite disasters.

Undernutrition, a chief child killer in developing countries, has been a major public health problem in Haiti. Following the 2010 disasters (earthquake and cholera) and the intensive relief efforts to address them, we sought to determine the trends of child undernutrition in Haiti using data from the 2005-06 Haiti Demographic and Health Survey (HDHS) and from a Standardized Monitoring and Assessment of Relief and Transitions (SMART) survey in 2012. Growth data analyses included 2,463 (HDHS) and 4,727 (SMART) children ages 0-59 months. We calculated the prevalence of stunting, wasting, and underweight for each survey using World Health Organization 2006 growth standards. To account for sampling design, probability weights were applied to all analyses. Statistical significance was determined by non-overlapping confidence intervals around estimates. Stunting prevalence declined from 28.5% (95% confidence interval [CI] = 25.9, 31.3) in 2005-06 to 22.2% (95% CI = 20.2, 24.3) in 2012; wasting, from 10.1% (95% CI = 8.2, 12.7) to 4.3% (95% CI = 3.6, 5.2); and underweight, from 17.7 % (95% CI = 15.6, 20.1) to 10.5% (95% CI = 9.3, 11.9). Additionally, stunting declined more in rural areas, from 33.6% (95% CI = 30.1, 37.2) in 2005-06 to 25% (95% CI = 23.4, 26.7) in 2012, than in urban areas, from 18.6% (95% CI = 15.3, 22.5) in 2005-06 to 18.4% (95% CI = 16.7, 20.1) in 2012, for reasons that remain unknown. Results of the 2012 HDHS confirmed the observed trends. Thus, undernutrition among Haitian children under 5 declined significantly between 2005-06 and 2012. Our results should be interpreted in view of investments and changes that occurred in different sectors (within and outside health and nutrition) before and after the earthquake.

Multilevel modelling of sociodemographic predictors of various levels of anaemia among women in Mali.

OBJECTIVE: Anaemia currently affects 40-80% of women in Africa. While risk factors for anaemia have been intensively studied, research has rarely compared risk factors between mild anaemia and moderate/severe anaemia. Also, the contribution of neighbourhood to the prevalence of anaemia has been rarely studied. The aim of the present study was to identify and compare individual and contextual factors associated with various levels of anaemia among women. DESIGN: A multilevel analysis of data from the 2001 Mali Demographic and Health Survey (n 3763) was carried out. Outcomes variables were mild anaemia, moderate-to-severe anaemia and any anaemia. Multilevel regression analyses were performed for each outcome. SETTING: Mali, West Africa. SUBJECTS: Women (n 3763) aged 15 to 49 years, including 512 pregnant women. RESULTS: Among the eleven potential risk factors included in the models, two factors were associated with mild anaemia (BMI and education), three with any anaemia (pregnancy, BMI and education) and six with moderate-to-severe anaemia (pregnancy, BMI, education, wealth, childhood residency and region of residence). Clustering of anaemia within communities was 20% for moderate-to-severe anaemia and 13% for mild anaemia. Despite significant differences in the prevalence of anaemia across regions in Mali, no difference between regions in the risk of mild anaemia was found and only the region of Gao showed a significantly higher risk of moderate-to-severe anaemia. CONCLUSIONS: The findings indicate that sociodemographic risk factors as well as clustering of anaemia varies with the severity of anaemia. Specific studies are needed to identify risk factors of mild anaemia as well as its consequences, as mild anaemia accounts for 20-40% of total prevalence of anaemia in Africa.

Risk factors for moderate to severe anemia among children in Benin and Mali: insights from a multilevel analysis.

BACKGROUND: Anemia currently affects 2 billion people throughout the world. Although the immediate causes of anemia among children are known (including malnutrition and infections), the importance of contextual determinants and their relationships with individual effects have rarely been explored. OBJECTIVE: To identify anemia risk factors at the individual, household, and community levels among Beninese and Malian children, using simple and multilevel regression methods. METHODS: An analysis was undertaken of nationally representative data collected in 2001 in Benin (n = 2,284) and Mali (n = 2,826) by the Demographic and Health Surveys. Sixteen potential risk factors for anemia were considered at the individual, household, and community levels. Comparative analyses were carried out using simple and multilevel logistic regression models. RESULTS: Simple and multilevel logistic regression analyses yielded broadly similar results. Risk factors for moderate to severe anemia included incomplete immunization, stunted growth, recent infection, absence of bednet, low household living standard, rural residency (Mali), low maternal education, and low community development index (Benin). In addition, multilevel analysis indicated a clustering level of anemia in communities (intraclass correlation) of 14% and 19% in Benin and Mali, respectively. CONCLUSIONS: Risk factors for child anemia appeared at all three levels (individual, household and community). Community-level clustering seemed to be low. Therefore, interventions to address anemia need not be village- or region-specific. Identifying a successful and replicable program is now a priority in child survival endeavors. It is likely that such a program would include a focus on improving immunization coverage, increased bednet usage, and reduced protein-energy malnutrition.