Harnessing New Computing Technologies for Detailed Assessment of Child Diet in Rural Zambia.

Objectives: To integrate tablet-based data collection tools with dietary data processing software for assessment of dietary intakes among children in rural Zambia. Methods: Dietary intakes of 4- to 8-year-old children in rural Zambia (n=1226) were collected by 24 hour recall interview using a customized survey tool on Android tablets. The 24 hour recall tool was pre-programmed with interviewer prompts, contained 1205 locally consumed foods, and collected detailed food descriptions including portion size, cooking method, added ingredients and source. Interview data were uploaded from the tablets to secure servers daily. Compiled data will be formatted for import into CS Dietary software, a program developed by Harvest Plus for entry and analysis of 24 hour recall data. CS Dietary analyzes foods and nutrients based on food composition, recipe, measurement conversion and food group tables which are selected by the researcher to match the research context. Tables corresponding to the food list used in the tabletbased recall tool will be imported into CS Dietary to create reports on diets of the children surveyed. Results: The total number of foods recorded was 8418. Of these, 31% (n=2620) were stiff, white maize porridge, 21% (n=1739) were leafy vegetable dishes and 11% (n=939) were small, whole fish dishes. Other foods consumed included large fish, eggs, beans, fritters, rice, and other maize preparations. Conclusions: This use of a tablet-based survey tool and CS Dietary software enabled paperless data collection and analysis based on context-specific food and recipe tables and shows the way forward for the deployment of customizable, interview-to-intakes tools for dietary assessment.

Micronutrient Status during Pregnancy and Effects of Antenatal Supplementation with Multiple Micronutrients versus Iron-folic Acid in Rural Bangladesh.

Objectives: Women often enter pregnancy with micronutrient deficiencies, exacerbated by demands of pregnancy. Yet, gestational micronutrient status is uncommonly assessed, even as momentum builds toward preventing multiple micronutrient (MM) deficiencies. We assessed micronutrient status of women early in pregnancy in a rural setting in northern Bangladesh. Methods: In a substudy of women participating in a randomized trial of MM versus iron-folic acid (IFA) supplementation we identified a population-based subsample of n=1526 women from whom plasma samples were obtained in the 1st trimester (TM), prior to supplementation, and in the 3rd TM, to evaluate micronutrient status and response to supplementation. Results: In available 1st TM data (n=491-1448 per nutrient assay), 6.8% of women were vitamin A deficient (retinol < 0.70 μmol/L), 41.7% had low β-carotene (<0.09 μmol/L), and 57.7% were vitamin E deficient (α-tocopherol < 12 μmol/L). Folate deficiency (plasma folate < 6.8 nmol/L) was 2.6%, while prevalence of vitamin B12 deficiency (cobalamin < 150 pmol/L) was 29.8%. Anemia affected 20.6% of women, but iron deficiency by TfR (4.7% > 8.3 μg/L) and ferritin (1.8% < 12 μg/L) was uncommon. Plasma zinc was low (<10 μmol/L) in 30.6% of women. Inflammation, by α-1 acid glycoprotein (AGP > 1 g/L), was present in 8.1% of women. Baseline vitamin D and iodine status, and MM versus IFA effects on micronutrient status by the 3rd TM, are being determined. Conclusions: Women in rural Bangladesh experience a variety of micronutrient deficiencies in early pregnancy, supporting the need to address "hidden hunger" with multiple micronutrient supplementation during pregnancy.

Malaria Exacerbates Inflammation-induced Bias in Ferritin and Soluble Transferrin Receptor Values.

Objectives: In the context of malaria and inflammation, the utility of ferritin and soluble transferring receptor (sTfR), as indicators of iron status may be compromised. In this study, we evaluated the effects of correcting for malaria and inflammation on the prevalence of iron deficiency (ID) as estimated by a) ferritin and b) sTfR. Methods: The analyses used baseline data from 1085 children, 4-8 y, who participated in a carotenoid biofortified maize flour trial in rural Zambia. For each biomarker, we compared the prevalence of ID with the prevalence corrected for a) CRP and AGP only; and b) CRP, AGP and concurrent malaria. Inflammation was defined as CRP>5mg/L and/or AGP>1g/L. Malaria was defined by microscopy. Children were first stratified into groups defined by inflammation and malaria status. Correction factors were then generated by dividing the group geometric means by that of the reference group (those free of both malaria and inflammation). Correction factors were applied to each individual concentration to generated corrected concentrations. Results: For ferritin, the unadjusted prevalence of ID (WHO age-specific cut-offs) increased from 7.3% to 9.5% (p<0.01) and 10.3 %( p<0.01), respectively, after correcting for CRP/AGP only, and CRP, AGP and concurrent malaria combined. For sTfR, the unadjusted ID prevalence (cutoff >8.3 mg/l) decreased from 28% to 21% (p<0.01) after correcting CRP/AGP only, and 19% (p<0.01) after correcting for CRP, AGP and concurrent malaria. Conclusions: Our findings highlight the need to account for both malaria and inflammation when interpreting ferritin and sTfr concentrations in malaria endemic regions.

Risk Factors for Anemia in Non-pregnant Women across the Agroecological Zones of Nepal.

Objectives: Describe the prevalence of anemia in Nepali non-pregnant women of reproductive age by agroecological zone and potential risk factors. Methods: We randomly sampled 21 Village Development Committees, each with 3 wards, proportionate to size across the mountains, hills and terai. In 4509 households with children <5 years or newlyweds we assessed anthropometry, 7-day diet frequency, morbidity, participation in health services and altitude-adjusted hemoglobin measurement (n=887). Household SES and empowerment questions were asked. Regression analyses were used to evaluate risk factors of anemia. Results: Anemia affected 53.3% of all women, and 42.3%, 35.8% and 66.1% in the mountains, hills and terai, respectively. Lowest vs. higher castes in the terai were at higher risk of anemia (OR=1.58, CI: 1.04-2.40). A secondary education (6-9 years) and overweightness (BMI > 24.9) appeared protective against anemia (OR=0.59, CI: 0.39-0.90 and OR=0.49, CI: 0.27-0.90, respectively). Women who accessed skilled health care workers in the past year had an increased risk of anemia (OR=1.36, CI: 0.99-1.85), likely reflecting their illness. Knowledge of deworming and iron-folic acid supplementation, SES and diversity of recent diet were unrelated to anemia risk. Conclusions: Approximately half of women of reproductive age have anemia in Nepal, with the highest risk in the terai. Anemia was less likely in educated, higher caste and better nourished women. Lack of association with recent diet may reflect inadequate duration of assessment with 7- day frequencies or causes of anemia other than iron deficiency. Health care providers are more likely to see anemic women, representing an opportunity to screen and treat anemia.

Testing Feasibility of Delivering Vitamin A to Newborns in Bangladesh’s Health Care System.

Objectives: Supplementing newborns with vitamin A within 48 hours of birth has been shown to reduce infant mortality in Bangladesh, India and Indonesia. This study evaluated the feasibility of delivering vitamin A to newborns through existing maternal and child health services in Bangladesh. Methods: Two sub-districts in each of 3 districts were randomized to one of two newborn vitamin A supplement (NVAS)distribution strategies: (1) providing theNVASandadministration instructions to pregnant women during clinic-based ANC visits (ANC model), and (2) having a health worker directly dose newborns during a post-natal visit (PNC model).All intervention components (i.e. health worker training, vitamin A capsule supplies, delivery of education messages and birth notification) were implemented through the existing health infrastructure. Primary outcomes, assessed by a pre-post survey, included coverage (% of newborns dosed) and timing of dose (% dosed within 48 hrs of birth) in December 2010 and June 2011. Results: 54% and 40% of newborns were supplemented in ANC and PNC models, respectively. 75%of dosed infants were reached within 48 hours of birth. Integrating NVAS into the health program did not increase ANC coverage, but significantly increased post-natal visits in both models(from 38% to 67% and 30% to 62% in the ANC and PNC models, respectively). Newborn VAS dosing was positively perceived by health workers and mothers. There were no serious side effects.

Dietary Diversity is Low and Associated with Nutritional Status of Women in Early Pregnancy in Rural Bangladesh.

Objectives: To evaluate the association between nutritional status and other risk factors with dietary diversity in the 1st trimester of pregnancy among women in rural Bangladesh. Methods: A cross-sectional analysis of data among women enrolled during the 1st trimester of pregnancy into the JiVitA-1 weekly vitamin A or beta-carotene supplementation trial in rural Bangladesh from 2001 to 2007. A dietary diversity score (DDS) based on 9-conventionally defined food groups was calculated based on data collected from a 7-day food frequency questionnaire. Mid-upper arm circumference (MUAC) from the main trial (n=67,690), and BMI and serum markers of micronutrient status from a substudy (n=1,869) of women were used as indicators of maternal nutritional status. Other individual, household and seasonality factors were examined for their association with DDS, and as potential confounders in the association between DDS and maternal nutritional status. Results: Overall mean (sd) DDS was low at 2.1 (1.1) (Max. Score=9). In a multivariate model, selected individual (maternal education, being a wage earner, and selected morbidity symptoms), household (living standard index, smaller household size, food security, and ownership of a fruit grove, home garden or fish pond) factors and seasonality (non-lean season) were positively associated with maternal DDS. Maternal MUAC, BMI and serum levels of lycopene (biomarker for vegetable/fruit intake) were also associated with DDS. Conclusions: In this rural setting of northern Bangladesh, dietary diversity is low and maternal nutritional status is poor. Dietary diversity, measured by a simple score summed over one week, was positively associated with maternal nutritional status in early pregnancy.

Constructing Indices of Rural Living Standards in Northwestern Bangladesh.

This study aimed to construct indices of living standards in rural Bangladesh that could be useful to study health outcomes or identify target populations for poverty-alleviation programmes. The indices were constructed using principal component analysis of data on household assets and house construction materials. Their robustness and use was tested and found to be internally consistent and correlated with maternal and infant health, nutritional and demographic indicators, and infant mortality. Indices derived from 9 or 10 household asset variables performed well; little was gained by adding more variables but problems emerged if fewer variables were used. A ranking of the most informative assets from this rural, South Asian context is provided. Living standards consistently and significantly improved over the six-year study period. It is concluded that simple household socioeconomic data, collected under field conditions, can be used for constructing reliable and useful indices of living standards in rural South Asian communities that can assist in the assessment of health, quality of life, and capabilities of households and their members.

Development and management of a geographic information system for health research in a developing-country setting: a case study from Bangladesh.

In the last decade, geographic information systems (GIS) have become accessible to researchers in developing countries, yet guidance remains sparse for developing a GIS. Drawing on experience in developing a GIS for a large community trial in rural Bangladesh, six stages for constructing, maintaining, and using a GIS for health research purposes were outlined. The system contains 0.25 million landmarks, including 150,000 houses, in an area of 435 sq km with over 650,000 people. Assuming access to reasonably accurate paper boundary maps of the intended working area and the absence of pre-existing digital local-area maps, the six stages are: to (a) digitize and update existing paper maps, (b) join the digitized maps into a large-area map, (c) reference this large-area map to a geographic coordinate system, (d) insert location landmarks of interest, (e) maintain the GIS, and (f) link it to other research databases. These basic steps can produce a household-level, updated, scaleable GIS that can both enhance field efficiency and support epidemiologic analyses of demographic patterns, diseases, and health outcomes.