![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Department of Nutritional Sciences, University of Connecticut, Storrs, CT 06269-4017 and * Department of Social and Preventive Medicine, University of Campinas, São Paulo, Brazil
2To whom correspondence should be addressed. E-mail: rafael.perez-escamilla{at}uconn.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT SUBJECTS AND PROCEDURESRESULTSDISCUSSIONLITERATURE CITED |
|---|
was 0.91 and the scale item response curves were parallel across the 4 household income strata. FI severity level was strongly associated in a dose-response manner (P < 0.001) with income strata and the probability of daily intake of fruits, vegetables, meat/fish, and dairy. These findings were replicated in the 2 independent survey samples. Results indicate that the adapted version of the USDA food insecurity module is valid for the population of Campinas. This validation methodology has now been replicated in urban and/or rural areas of 4 additional states with similar results. Thus, Brazil now has a household food insecurity instrument that can be used to set national goals, to follow progress, and to evaluate its national hunger and poverty eradication programs.
KEY WORDS: • Brazil • household food insecurity • Fome Zero Program • hunger • USDA food insecurity module
According to both the USDA and the WHO, household food security can be defined as access to a diet of enough quantity and quality for all household members at all times and through socially acceptable ways to maximize the chances for a healthy and active life. Food insecurity (FI)3 has been assessed indirectly through energy balance sheets and child anthropometry. However, these 2 approaches have not always been useful for guiding food security polices at the national, regional, or local level (1,2). Thus, researchers recognized the need to also measure this phenomenon through more direct experiential approaches at the household level. During the 1990s, the USDA led the effort to develop a valid scale that was capable of measuring household food insecurity in the United States (3,4). This work built heavily on the Radimer/Cornell hunger scale (5,6) and the Childhood Hunger Identification Project scale (7). The USDA-led effort eventually resulted in the adoption of a standard FI module in the U.S. Current Population Survey as well as in the National Health and Nutrition Examination Survey. The availability of this tool fostered an exponential growth in U.S. FI prevalence surveys and research seeking to understand both the FI determinants and its consequences (3,5,8–11). Because of the simplicity of the USDA scale, several countries expressed an interest in adopting it for assessing household FI in their populations. The case in point is Brazil, a country with a hunger eradication program in place known as "Fome Zero" ("Zero Hunger"). In spite of this national agenda, until recently, Brazil had not adopted a valid and standard FI instrument that would allow the government to set FI national goals, improve the targeting of their food and social assistance programs, and measure progress across time. Thus, the objectives of this study were as follows: 1) to describe the process of adaptation and validation of the USDA FI scale, and 2) to assess the validity of this scale in an urban area in Brazil.
|
SUBJECTS AND PROCEDURES |
|---|
|
TOPABSTRACTSUBJECTS AND PROCEDURESRESULTSDISCUSSIONLITERATURE CITED |
|---|
Setting.
Campinas has 
1 million inhabitants and is located 100 km northwest of the City of São Paulo. The city, originally founded by coffee barons, developed as one of the major industrial corridors in Brazil, housing major transnational companies in the chemical, computer, and biotechnology sectors. Although some of the neighborhoods in this city are very wealthy, others are extremely impoverished shanty towns or "favelas." Thus, Campinas is an ideal setting for testing the psychometric behavior of the food insecurity scale.
Instrument adaptation and qualitative validation. The 15-item USDA FI scale and its 3 subitems were translated into Portuguese by one of the authors (A.M.S.-C.). The translated instrument was then subjected to a question-by-question review by a panel of 13 experts on FI and/or public health nutrition from the University of Campinas, the Health and Welfare Secretary of the district of Campinas, the Ministry of Health, the "Fome Zero" coordinator in Campinas, and the University of Connecticut. The revised instrument that resulted from this meeting was then presented and thoroughly discussed at a focus group meeting with community members who had experienced household food insecurity; the meeting took place in a local church in the City’s southwest health district. The focus groups participants’ selection process also ensured that a range of ages and both genders were represented. This meeting was attended by 3 adult women and 3 adult men ranging from young to elderly adults and was moderated by a research staff member with expertise in anthropological research. The 2 study directors, as well as a local nutritionist, were also present and took notes throughout the session. Participants were not offered any economic incentive for their participation and were clearly told that the information provided would not be used in any decision concerning food assistance or social benefits. The focus group meeting lasted 3 h and was divided into 2 segments. First, participants were asked to provide their definitions of the following terms each of which was prominently displayed, one at a time, in a paper sheet placed in the center of the circle: "varied diet," "healthy diet," "healthy and varied diet," "dietary quality," "enough food," "enough money," "nutritious food," "hunger," "food security," and "food insecurity." The 2nd segment of the focus group involved a question-by-question discussion of the modified instrument resulting from the expert panel meeting. Soon after the convenience sample survey described below, a 3rd focus group was conducted with the interviewers participating in the quantitative phase of the scale validity assessment to confirm the understanding of the questionnaire items by respondents and the ease of application of the instrument.
Instrument quantitative validation. The adapted instrument resulting from the focus groups (Table 1) was then used for the quantitative validation phase, which included both a convenience and a representative sample from the City of Campinas.4
|
Representative survey. Once the quantitative validity of the survey was established, the FI scale was applied to a representative sample of the city. A sample of the noninstitutionalized civilian population living in urban areas in the municipality of Campinas was selected using stratified cluster sampling in 2 stages. The census tracks correspond to the primary sampling units (PSUs) and the households to the secondary units. The sampling framework used was similar to the one used for the 2000 Health Survey of the State of São Paulo (ISA-SP). The PSUs were 30 census tracks selected with probability proportional to the number of households and drawn from 3 strata (percentage of head of household with college level education: 25 vs. 26–50 vs. > 50%) in equal numbers (i.e., 10 tracks per strata). This sampling framework was based on the 1996 census conducted by the Brasilian Institute of Geography and Statistics and an updated listing and mapping of all households in the 30 census tracks that took place in 2000 and was carried out by ISA-SP personnel.
Of the 1000 households randomly chosen, 847 responded to the interview, yielding a nonresponse rate of 15.3%. The overwhelming reason explaining the nonresponse rate was the inability to find these households. Interviews were conducted during weekend days to maximize the chances of finding the target respondent (i.e., person in charge of food preparation). When necessary, households were revisited until the target respondent was found. Interviews were conducted by trained and supervised college students in the fields of nutrition, nursing, and food and agricultural engineering. Interviewers were randomly assigned to the different census tracks and were provided with accurate maps and addresses of the randomly chosen households.
Household income.
For identifying social strata in the convenience sample, the survey included a monthly household income question from which respondents had to choose one of the following monthly income ranges, expressed as multiples of the official minimum wage in Brazil: <1 minimum wage, 1–2 minimum wages, 3–4 minimum wages, and 
5 minimum wages. The representative sample survey itemized all income sources available to the family and then classified the total monthly income into the same minimum wage strata.
Food intake. Food group intake was measured with a short food group frequency questionnaire, designed specifically for this study; the questionnaire asked the respondent whether s/he consumed every day (at least once per day) the following: cereals, tubers and roots, milk, dairy products, eggs, fruits and natural juices, vegetables, legumes, meat/poultry/fish, candy, soft drinks.
Validity criteria.
As recommended by Frongillo (12), there were 4 validity criteria established a priori: 1) an expected Chronbach 0.85; 2) parallelism on item response curves across socioeconomic strata; 3) a clear-cut dose-response relationship between socioeconomic strata and level of FI; 4) a clear-cut dose-response between "nutritious foods" (fruits, vegetables, animal protein, dairy) consumption and level of FI. In addition, we added a 5th validity criterion requiring that the replicability of findings in 2 independent samples drawn from the same population be demonstrated.
Data analyses.
SPSS for Windows (version 11.5) was used to enter the data and conduct all statistical analyses for the convenience sample survey. Analyses and estimates from the representative survey were conducted taking into account the complex survey sampling design using the Complex Samples module from SPSS for Windows (version 12.0). Data were entered and cleaned by trained and supervised graduate students at the Department of Social and Preventive Medicine at UNICAMP. All of the FI scale questions were recoded into 2 categories [(often/sometimes vs. never) or (yes vs. no)]. Each item was given a score of 1 if the answer pointed toward FI, or 0 if it was in the food security direction. Among convenience sample respondents who gave a negative answer to the first 4 items, the likelihood for giving a positive response to any of the subsequent items was negligible. Thus, for the representative sample survey, it was decided not to ask the rest of the items among respondents answering negatively to the first 4 scale items and to assume in the analyses that their responses to these items was also negative. The subquestions related to frequency of occurrence were not included in the analyses. Thus, the theoretical food insecurity score range was from 0 to 15 in households with children/teenagers and from 0 to 9 in households in which only adults lived. An additive total score was created and households were classified into 4 mutually exclusive levels of food (in)security using the following "equidistant" algorithms: 1) food secure (score = 0); 2) mild FI [score: 1–5 (households with children/teenagers; 1–3 (households with only adults)]; 3) moderate FI (6–10; 4–6); and 4) severe FI (11–15; 7–9). The Chronbach internal consistency test was run with households with children/adolescents as a basis because these were the ones with responses available for all 15 items.
To test the parallelism of the item response curves, we plotted the percentage of affirmative responses to each of the FI scale items across 4 household income strata. The 
2-square test was used to examine the association between household income and items’ response (yes/no), and the associations of FI severity category with food intake and socioeconomic strata. Statistical significance was based on a 2-sided probability value 
0.05.
|
RESULTS |
|---|
|
TOPABSTRACTSUBJECTS AND PROCEDURESRESULTSDISCUSSIONLITERATURE CITED |
|---|
50% of them were Caucasian. Although there was 1 more room per household in the representative than in the convenience sample, the number of bedrooms was similar. Given the sampling approach used with the convenience sample, it is not surprising that households participating in that survey were more likely than households in the representative survey to have member(s) 18 y old. Similarly, it is not surprising that households from the convenience sample were poorer(Table 2).
|
Furthermore, the USDA food (in)security questionnaire was rated as highly appropriate by both the content experts and the community members participating in the focus groups. However, the following modifications were recommended and implemented (Table 1): 1) questions were asked in relation to the 3 mo, rather than the 12 mo preceding the survey; 2) questions targeting households with member(s) 18 y used the term "children/adolescents" instead of just children; 3) the term "healthy and varied" diet replaced the term "balanced meal" used in the original USDA FI module; 4) all questionnaire items were constructed as questions rather than statements; 5) each question receiving an affirmative response was followed by a subquestion assessing to what extent the FI problem was experienced; and 6) at the beginning of the survey, it was made clear that this information was not going to be used to either include or exclude anybody from food and/or social assistance programs.
The clarity of the food insecurity module was further confirmed at the focus group with the interviewers who applied the convenience sample survey. Thus, minimal changes had to be implemented to the FI module subsequently used in the representative survey.
Internal consistency and parallelism.
Chronbach’s was 0.91 in both the convenience (n = 119, 15 items) and the representative (n = 456, 15 items) sample. Except for the last item in the representative sample survey, the scale item response curves in both surveys were parallel across income strata indicating that the likelihood of an affirmative response to all items increased as monthly household income decreased (Fig. 1).
|
|
Food insecurity severity level and food intake. The analyses detected foods that were highly sensitive (Fig. 3) and foods that were insensitive or only moderately sensitive (Fig. 4) to the level of food insecurity. Figure 3 shows that in both survey samples, the level of food insecurity was significantly (P < 0.001) and strongly associated with the likelihood of daily consumption of fruit, non-root/tuber vegetables, and meat. Among convenience sample households that were severely food insecure, the likelihood of daily consumption of fruits, non-root/tuber vegetables, and meat was zero. Among representative sample participants, this likelihood ranged from 13.5 to 31.9%. By contrast, the likelihood of daily consumption of these foods was very high among food-secure households in both surveys, ranging from 70.4 to 91%. Similar patterns of association were found for milk, dairy products, juices, candy, and soft drinks (data not shown).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTSUBJECTS AND PROCEDURESRESULTSDISCUSSIONLITERATURE CITED |
|---|
The USDA food insecurity module or primary scales from which it was developed were used to assess food insecurity determinants or consequences in countries as diverse as Indonesia (13) and Venezuela (14), and among different ethnic groups in the U.S. and Canada (12), such as Latinos (9) and Asian-Pacific islanders (15) not specifically included in the original development and validation. In the few instances in which rigorous psychometric testing was conducted (12,14,15), results are fully consistent with our conclusion that the USDA FI module has a strong external validity.
Key adaptations deemed necessary for the application of the USDA FI module in Campinas are highly consistent with qualitative results reported by Harrison et al. (16) when they tested a translated version of the instrument with Latinos living in the U.S. In both instances, the term "balanced meal" was found to be difficult to interpret by most community members. Similarly, they also concluded that the grammatical structure had to be simplified. An important qualitative finding in our study was the strong association that focus group participants made between dietary quality and the microbial and overall safety of foods (i.e., foods that don’t make you sick). Thus, it is important that future research efforts identify valid food safety questions that can be incorporated into household FI modules.
It is important to recognize that researchers who applied a qualitative methodology similar to the one used for the development of the Radimer/Cornell hunger scale in South Asia, ended up proposing food insecurity scales that differed substantially from the USDA food insecurity module (17). Provided that these efforts are followed by rigorous psychometric validations, these approaches are invaluable for specific local research and/or programmatic efforts and complement enormously the approaches needed to meet the need for "national" food insecurity scales that can be used to set national goals and follow trends.
A unique contribution from this study is its multi-institutional and community participatory nature, and the fact that it used both qualitative and quantitative approaches to test the validity of the scale in a developing country. As a result, the government of Brazil welcomed the results from this rapid validation assessment. Thus, the effort initiated in Campinas was extended nationally and has now involved the replication of the validation methodology in 3 additional urban areas in the Sates of Amazonas, Paraíba, and Brasília and in 5 rural areas including the 4 original states and the State of Mato Grosso. Results from these replications were totally consistent with the Campinas findings (18); as a result, the Brazilian government decided to incorporate the adapted FI module into national surveys and to make it available to researchers all over the country. This was an essential step that Brazil had to take to set national goals and be able to monitor the effect of its "Fome Zero" policies, which represent a top national priority for its government.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
3 Abbreviations used: FI, food insecurity; ISA-SP, Health Survey of the State of São Paulo; PSU, primary sampling unit.
4 The Portuguese version of Table 1, the food insecurity questionnaire items, is available with the on-line posting of this paper at www.nutrition.org.
Manuscript received 23 April 2004. Initial review completed 11 May 2004. Revision accepted 25 May 2004.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTSUBJECTS AND PROCEDURESRESULTSDISCUSSIONLITERATURE CITED |
|---|
1. Pelletier, D. L., Deneke, K, Kidane, Y., Haile, B & Negussie, F. (1995) The food-first bias and nutrition policy: lessons from Ethiopia. Food Policy 20:279-298.[Medline]
2. Mulder-Sibanda, M., Sibanda-Mulder, F. S., d’Alois, L. & Verna, D. (2002) Malnutrition in food-surplus areas: experience from nutritional surveillance for decentralized planning in Haiti. Food Nutr. Bull. 23:253-261.[Medline]
3. Carlson, S. J., Andrews, M. S. & Bickel, G. W. (1999) Measuring food insecurity and hunger in the United States: development of a national benchmark measure and prevalence estimates. J. Nutr. 129:510S-516S.
4. Rose, D., Basiotis, P. P. & Klein, B. W. (1995) Improving federal efforts to assess hunger and food insecurity. Food Rev. 18:18-23.
5. Radimer, K. L., Olson, C. M., Greene, J. C., Campbell, C. C. & Habicht, J. P. (1992) Understanding hunger and developing items to assess it in women and children. J. Nutr. Ed. 24:36S-44S.
6. Radimer, K. L. (2002) Measurement of household food security in the USA and other industrialized countries. Public Health Nutr. 5:859-864.[Medline]
7. Wehler, C. A., Scott, R. I. & Anderson, J. J. (1992) The community childhood identification project: a model of domestic hunger–demonstration project in Seattle, Washington. J. Nutr. Ed. 24:29S-35S.
8. Derrickson, J. P., Fisher, A. G., Anderson, J. E. & Brown, A. C. (2001) An assessment of various household food security measures in Hawaii has implications for national food security research and monitoring. J. Nutr. 131:749-757.
9. Pérez-Escamilla, R., Ferris, A. M., Drake, L., Haldeman, L., Peranick, J., Campbell, M., Peng, Y. K., Burke, G. & Bernstein, B. (2000) Food stamps are associated with food security and dietary intake of inner-city preschoolers from Hartford, Connecticut. J. Nutr. 130:2711-2717.
10. Nord, M., Andrews, M. & Winicki, J. (2002) Frequency and duration of food insecurity and hunger in US households. J. Nutr. Educ. Behav. 34:194-200.[Medline]
11. Rose, D. (1999) Economic determinants and dietary consequences of food insecurity in the United States. J. Nutr. 129:517S-520S.
12. Frongillo, E. A. (1999) Validation of measures of food insecurity and hunger. J. Nutr. 129:506S-509S.
13. Studdert, L. J., Frongillo, E. A., Jr & Valois, P. (2001) Household food insecurity was prevalent in Java during Indonesia’s economic crisis. J. Nutr. 131:2685-2891.
14. Albert, P. L. & Sanjur, D. (2000) The adaptation and validation of the food security scale in a community of Caracas, Venezuela. Arch. Latinoam. Nutr. 50:334-340.[Medline]
15. Derrickson, J. P., Fisher, A. G. & Anderson, J. E. (2000) The core food security module scale measure is valid and reliable when used with Asians and Pacific Islanders. J. Nutr. 130:2666-2674.
16. Harrison, G. G., Stormer, A., Herman, D. R. & Winham, D. M. (2003) Development of a Spanish-language version of the U.S. household food security survey module. J. Nutr. 133:1192-1197.
17. Frongillo, E. A., Chowdhury, N., Ekstrom, E. C. & Naved, R. T. (2003) Understanding the experience of household food insecurity in rural Bangladesh leads to a measure different from that used in other countries. J. Nutr. 133:4158-4162.
18. Segall Corrêa, A. M., Pérez-Escamilla, R., Kurmaran Maranha, L., Sampaio, F., Vianna, R., Vieira-Freire, L., Yuyama, L., Alencar, F., Coitinho, D., Gubert, M., Leão, M., Schimtiz, B., Portela de Abuquerque, Z. & Escamilla, J. A. (2004) An adapted version of the USDA food security scale is a valid tool to assess household food insecurity in urban areas in Brazil. FASEB J. 18:A488 (abs.).
This article has been cited by other articles:
|
|
 |
|
  R. Perez-Escamilla, M. Dessalines, M. Finnigan, H. Pachon, A. Hromi-Fiedler, and N. Gupta Household Food Insecurity Is Associated with Childhood Malaria in Rural Haiti J. Nutr., November 1, 2009; 139(11): 2132 - 2138. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. C. Maes, C. Hadley, F. Tesfaye, S. Shifferaw, and Y. A. Tesfaye Food Insecurity among Volunteer AIDS Caregivers in Addis Ababa, Ethiopia Was Highly Prevalent but Buffered from the 2008 Food Crisis J. Nutr., September 1, 2009; 139(9): 1758 - 1764. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Hackett, H. Melgar-Quinonez, R. Perez-Escamilla, and A. M. Segall-Correa Gender of respondent does not affect the psychometric properties of the Brazilian Household Food Security Scale Int. J. Epidemiol., August 1, 2008; 37(4): 766 - 774. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Gonzalez, A. Jimenez, G. Madrigal, L. M. Munoz, and E. A. Frongillo Development and Validation of Measure of Household Food Insecurity in Urban Costa Rica Confirms Proposed Generic Questionnaire J. Nutr., March 1, 2008; 138(3): 587 - 592. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Isanaka, M. Mora-Plazas, S. Lopez-Arana, A. Baylin, and E. Villamor Food Insecurity Is Highly Prevalent and Predicts Underweight but Not Overweight in Adults and School Children from Bogota, Colombia J. Nutr., December 1, 2007; 137(12): 2747 - 2755. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. R. Melgar-Quinonez, A. C. Zubieta, B. MkNelly, A. Nteziyaremye, M. F. D. Gerardo, and C. Dunford Household Food Insecurity and Food Expenditure in Bolivia, Burkina Faso, and the Philippines J. Nutr., May 1, 2006; 136(5): 1431S - 1437S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Coates, E. A. Frongillo, B. L. Rogers, P. Webb, P. E. Wilde, and R. Houser Commonalities in the Experience of Household Food Insecurity across Cultures: What Are Measures Missing? J. Nutr., May 1, 2006; 136(5): 1438S - 1448S. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
