1. Tufts University Friedman School of Nutrition Science and Policy
Household Food Security and the Nutritional Status of Rural Tanzanian Adolescents (10-19 years)
This study was conducted by the primary author, Lorraine Cordeiro, as part of her dissertation research under the auspices of UNICEF/Tanzania and Tufts University Friedman School of Nutrition Science and Policy.
Cordeiro LS, Wilde P, Pinderhughes E, Lamstein S, and Levinson FJ
Tufts University Friedman School of Nutrition Science and Policy

2. Abstract
Household food insecurity contributes to poor nutritional health, with negative consequences on growth and development across childhood. While the nutritional status of children under five years is a priority due to vulnerability and child survival concerns, once beyond that critical age, adolescence is the next important nutritional phase – a time when adult bodies, minds, and social behaviors are formed. This study investigates the association between household food insecurity and undernutrition (BMI for Age <5th percentile of the NCHS/WHO reference) among a sample of never-married adolescents ages years (n=670) from 28 villages in Kilosa District, Tanzania. This study is unique in examining these issues in the context of a developing country, Tanzania, which is deeply affected by HIV/AIDS and where chronic undernutrition is highly prevalent. The prevalence of undernutrition among adolescents in this sample was 21%. Multivariate analyses tested for associations between undernutrition and three distinct measures of household food insecurity (i.e. household caloric adequacy, household dietary diversity score, and coping strategies index), after adjusting for morbidity, socioeconomic status, and potential confounders. Household Dietary Diversity Score emerged the most important indicator of food insecurity: every one additional food group added to the household diet increased the odds of being well-nourished by 18% (OR 0.82). Similarly, the odds of being well-nourished increased by 9% (OR 0.91) with every unit increase (1 unit = 10 percentage points) in household caloric adequacy. These findings support a growing body of research on adolescent health, suggesting detrimental effects of household food insecurity on nutritional status. We conclude that household dietary diversity is a practical indicator of food security and a strong determinant of adolescent nutritional health, and recommend its use in studies on adolescents in developing countries.
Cordeiro et al. 2007

3. Cordeiro et al. 2007

4. Adolescents
Adolescents are defined as persons aged years (WHO 1986)
Over 20% of the world’s population (UNICEF 2000).
85% live in developing countries (Rosen 2004; UN 1997).
25% of Tanzania’s population (NBS 2002).
Adolescence is a time of experiential learning and vulnerability to environmental factors make adolescents particularly prone to engaging in sexual and other risk behaviors (Reininger 2003).
Globally, the highest prevalence of sexually transmitted infections (STIs), including HIV, is observed among individuals aged 15 and 24 years (WHO 2005)
Adolescence is an optimal time to establish healthy behaviors.
Cordeiro et al. 2007

5. Sub Saharan Africa
Youth are projected to comprise 33% of the total population in 2025
Undernutrition affects between 15-29% of adolescents
Young people (15-24 yrs) comprise more than 50% of new HIV infections (UNAIDS 2004)
Food security remains one of the continent’s most persistent challenges to economic growth
Cordeiro et al. 2007

6. Adverse Consequences of Household Food Insecurity on Adolescent Nutritional Status
Limited research on relationship between food insecurity and adolescent nutritional status
U.S. studies
links between FIS and obesity (Casey et al. 2001)
FIS negatively associated with the children’s BMIs and household food supplies but not with the children’s food intakes (Matheson et al. 2002)
Household food insecurity or insufficiency has been associated with poor child health in the U.S., with increased health risks particularly for stomach aches, headaches, depression, and suicide.
Cordeiro et al. 2007

7. Defining Terms Household Food Security
Food Security “exists when all people, at all times, have physical and economic access to sufficient, safe, and nutritious food to meet their dietary needs and food preferences for an active and healthy life” (World Food Summit 1996)
Cordeiro et al. 2007

8. Measures of Household Food Security
Household Caloric Adequacy per Adult Equivalent (CalAdq)
Household Met 80% of its Calorie Requirements per Adult Equivalent (Met80%) 
Household Diet Diversity Score (HDDS) 
Coping Strategies Index (CSI)
Due to the difficultly in assessing household food security status, four distinct measures were used in this study. These measures were correlated with undernutrition, assets, and income status, as well as with each other. The measures used were:
1) HH Caloric Adequacy per Adult Equivalent: developed using one 24-hour dietary recall per household. Raw and cooked food quantities were converted into nutrient quantities using the Tanzania Food Composition Table (Lukmanji et al. 2006). Adult Equivalencies were established using methodology proposed by Swindale and Ohri-pati (1999, 2006 FANTA).
2) Caloric Requirements per Adult Equivalent (Household Met 80%): This dichotomous variable, based on the household’s caloric adequacy, classified households that met less than 80% of their calorie requirements per adult equivalent as food insecure (coded ‘0’) and households that met or exceeded 80% of their calorie requirements as food secure (coded ‘1’).
3) Household Diet Diversity (HDDS), the number of different food groups represented in the diet, is often recommended as a proxy indicator of food security status (Hoddinott and Yohannes 2002; Coates et al. 2003; Torheim et al. 2004; Swindale and Bilinsky 2005). Using Swindale and Bilinsky’s (2005) methodology, 24-hour dietary recall data collected from each household was coded according to a set of 12 food groups and a household dietary diversity score was calculated based on this. HDDS was used a continuous variable in multivariate analyses and as a categorical variable in bivariate analyses (low diversity 1-3 food groups=0; average diversity 4-5 food groups=1; high diversity 6-9 food groups=2).
4) Coping Strategies Index (CSI) is a food security indicator that assesses coping strategies utilized by households during periods of food shortage (Maxwell 1999). It measures access to food, as well as both the frequency and severity of these coping strategies (Maxwell 1999). CSI is unique in that it incorporates vulnerability, an element of food insecurity that is often overlooked in conventional measures. The Coping Strategies module of the survey was developed using methodology detailed by Maxwell et al. (2002). This methodology incorporates the use of qualitative data. Extra information: In particular, focus group research conducted to formulate the Coping Strategies module proved to be useful in the interpretation of research findings. CSI was entered as a continuous variable in logistic regression models. A categorical variable, based on the number of strategies employed by households, was developed for descriptive purposes. Households that did not utilize any coping strategies were coded ‘0’, while those that used between 1-3 strategies were coded ‘1’, 4-6 strategies were coded ‘2’, and 7 or more strategies were coded ‘3’.
Cordeiro et al. 2007

9. Undernutrition
Body Mass Index (BMI=weight (kg)/height (m)2) for age and sex less than the 5th percentile of the NCHS/WHO reference population (WHO 1995b; Must et al. 1991).
Undernutrition was defined as body mass index (BMI=weight (kg)/height (m)2) for age and sex less than the 5th percentile of the NCHS/WHO reference population (WHO 1995b; Must et al. 1991).
For descriptive purposes, nutritional status was categorized into three groups in reference to the standard population: undernourished (<5th percentile of the standard), normal (5th ≤ 85th percentiles), and overweight (>85th percentile).
For multivariate analyses, respondents whose body mass indices for age and by gender fell below the 5th percentile of the standard were considered undernourished (coded ‘1’), while those with BMI for age by gender ≥ the 5th percentile were considered to be normal (coded ‘0’).
Cordeiro et al. 2007

10. Research Question What factors are associated with adolescent nutritional status in Tanzania?
Primary Hypothesis:
Household food insecurity is independently associated with undernutrition among adolescents after controlling for age, gender, orphan status, puberty, hygiene practices, morbidity, physical activity, socioeconomic status (i.e. assets and expenditure), and number of household members.
The primary hypothesis posed in this study. We controlled for several variables.
Cordeiro et al. 2007

11. Research Design and Data Collection
Cross-sectional Survey
January to March 2004
Sample:
30 villages (cluster sampling)
791 adolescents interviewed (simple random sample)
Response rate 94%
Unit of analysis: Adolescent (10-19 years)
Never-married Individuals (N=670)
The study was based on a cross-sectional design. Data were collected during the period of peak food insecurity (Jan-March) in this District of Tanzania.
Cluster sampling using population proportionate to size was utilized to select 30 villages in the district. We were able to successfully survey 28 villages and 2 villages were inaccessible due to heavy rains.
A total of 791 adolescents were interviewed, representing a high response rate of 94%. Rapport was built with village and community leader prior to survey implementation and households were informed of survey work 1 month prior to data collection.
Adolescents were randomly selected from a master list of all adolescents in the village – drawn from village registries.
This study presents findings from a sample of adolescents were never-married and between year at the time of data collection.
Cordeiro et al. 2007

12. Data Analysis Prevalence estimates and contigency tables
Correlation analysis
undernutrition, food security indicators, and measures of socioeconomic status.
Logistic regression models
tested the associations between undernutrition and each of the four measures of household food insecurity, after adjusting for age, gender, orphan status, puberty, hygiene practices, physical activity, morbidity, socioeconomic status, and number of household members.
Assessed whether or not food insecurity affected nutritional status through diminished health.
Several statistical methods were employed. The results presented today are prevalence estimates and multivariate analyses.
Cordeiro et al. 2007

13. Results
Table 1. Household Food Security Status Using Three Measures (N=637)
Results indicate high food insecurity in this population.
Of particular interest is the finding that more than 1/3 (36.9%) of households were consuming between 1-3 food groups. This represents the highest level of food insecurity and such percentages are expected primarily in complex emergency situations.
Cordeiro et al. 2007

14. Results
Table 2. Adolescent Health and Nutritional Status in Kilosa District (N=670)
Cordeiro et al. 2007

15. Results: Bivariate Analyses
A significantly higher proportion of male than female adolescents were undernourished (25.4% vs. 15.7%, p<0.01).
No Association was observed between morbidity and nutritional status in this sample.
The proportion of undernourished adolescents living in food insecure households was significantly higher than in food secure households (24.5% vs. 17.4%, p<0.05)
In general developing country males have been found to exhibit higher incidence of undernutrition than females. Further analyses of results indicate an association between nutritional status by gender and sexual maturation.
Morbidity was high (70%) and no assessment on the severity of morbidity was determined in this study.
As expected, a bivariate association between household food security status and nutritional status was observed.
Cordeiro et al. 2007

16. Results: Multivariate Analyses
Table 3. Undernutrition among Adolescents (10-19 yrs): A Comparison of Coefficients for Four Distinct Measures of Household Food Security
Coef. (SE)
P Value
Household Caloric Req. per Adult Equivalent (CalAdq)
(0.036)
0.009**
Household Met 80% of its Calorie Req. per Adult Equivalent (Met80%)
(0.251)
0.018*
Household Dietary Diversity Score (HDDS)
(0.094)
0.036*
Coping Strategies Index (CSI)
0.004 (0.380)
0.991
Logistic Regression Models control for age, gender, orphan status, pubertal status, poor hygiene practices, physical activity levels, morbidity, socioeconomic status, and number of household members
* P<0.05; **P<0.01; ***P<0.001
Cordeiro et al. 2007

17. Results: Multivariate Analyses
Undernutrition was significantly and negatively associated with household food insecurity after adjusting for potential confounding variables.
For every unit increase (1 unit=10 percentage points) in household caloric adequacy, the odds of being well-nourished increased by 9% (Odds Ratio (OR) 0.91) for adolescents in this sample.
Household Dietary Diversity Score: Consumption of every one additional food group at the household level increased the odds of an adolescent being well-nourished by 18% (OR 0.82), after controlling for all other covariates in the model.
Multivariate analyses:
Logistic regression techniques were applied to the data to determine the association between household food security status and undernutrition among adolescents in this population.
Our findings indicated that:
For every unit increase (1 unit=10 percentage points) in household caloric adequacy, the odds of being well-nourished increased by 9% (Odds Ratio (OR) 0.91) for adolescents in this sample.
and
Consumption of every one additional food group at the household level increased the odds of an adolescent being well-nourished by 18% (OR 0.82), after controlling for all other covariates in the model.
Cordeiro et al. 2007

18. Conclusions
Food insecurity is significantly and negatively associated with adolescent nutritional status.
Household dietary diversity is an acceptable and useful indicator of food security and a strong determinant of adolescent nutritional health.
Household dietary diversity might be the most practical indicator of food security in adolescent studies conducted in developing countries.
Cordeiro et al. 2007

19. Conclusions
The results of this study are promising as they suggest that simple strategies focused on increasing household caloric intake and improving dietary diversity among the most vulnerable households could improve the nutritional health of young people – even when they are not the intended beneficiaries of food aid and nutrition education strategies.
Cordeiro et al. 2007

20. Funding UNICEF/Tanzania Tufts University Center for Children
Helen Brownstein Scholarship Award
Ellen and Ronald Block Foundation Scholarship Award
Cordeiro et al. 2007

21. Research and programs carried out with a perspective that views children and adolescents as assets, incorporates youth as agents of change, and acknowledges the responsibility of adults in ensuring safe and healthy environments for child development hold the most promise in altering poor health among children and adolescents in Africa.
Cordeiro et al. 2007