Nutrient Intakes of Food Insufficient and Food Sufficient Adults in the Southern Region of the United States and the Impact of Federal Food Assistance Programs

Carol L. Connell, MS, RD
Kathy Yadrick, PhD, RD
Agnes W. Hinton, DrPH, RD
The University of Southern Mississippi
Hattiesburg, MS

L. Joseph Su, PhD, MPH
Louisiana State University Medical Center
New Orleans, LA

December 27, 2000

In order to understand the impact of food insufficiency and food/nutrition assistance program participation on persons residing in the South, this study examined the nature of food insufficiency among a Southern population using NHANES III data and the impact of food assistance programs on this phenomenon. Five research questions guided the data analysis.

1. Do food insufficient adults have significantly different nutrient intakes than food sufficient adults when controlling for other influences on nutrient intakes namely, education level, smoking status, age, gender of the household head, and body size?
2. What is the association between food insufficiency and nutrient intakes among these adults?
3. Are there significant differences in nutrient intakes of food insufficient adults based on participation in food assistance programs when controlling for other influences on nutrient intakes namely, education level, smoking status, age, gender of the household head, and body size?
4. Does the number of food assistance programs participated in by the family influence nutrient intake?
5. What is the association between participation in food/nutrition assistance programs and nutrient intakes among these adults?

Previous studies of the impact of food insecurity and hunger on food and nutrient intakes using both primary and secondary data have revealed lower intakes of several nutrients among women of child-bearing age, the elderly, poor Northeastern Caucasian women and low-income Canadian women. However, little has been done to date to try to define the food insecurity-related nutritional problems of specific regions of the U.S. such as the South. Factors unique to the South and to particular regions within the South warrant the investigation of the effects of food insecurity/insufficiency on nutrient intake in this region. In addition, no published studies have attempted to determine the impact food/nutrition assistance programs such as the FSP or WIC have on nutrient intakes of individuals from food insecure/insufficient households in the Southern region or the U.S.

The sample for this research was drawn from those adults, age 18 years and older, residing in the Southern region of the U.S. as defined by NHANES III. Adults were classified as food insufficient if the household food supply was reported as "sometimes" or "often" not enough to eat (n=456); the rest of the adults were classified as food-sufficient (n=6741). Least square means were calculated for each nutrient. Chi Square analysis was used to test for differences between groups based on demographic characteristics. Analysis of variance (ANOVA), adjusted for other factors influencing dietary intake, namely education level, smoking, age, gender of the household head, and body size, was performed to test the difference in nutrient intakes between food insufficient and food sufficient adults. This procedure was also used to test the difference between nutrient intakes of food insufficient participants and non-participants in food and nutrition assistance programs. Multiple regression analysis was used to examine the relationship between food insufficiency and nutrient intakes as well as the relationship between food assistance program participation and nutrient intakes, again accounting for potential confounding factors.

We found significant differences in demographic characteristics of the food sufficient compared to the food insufficient. Those most often reporting food insufficiency were either in the youngest age group, non-white, had low formal educational attainment, were in female-headed households or participated in only one food/nutrition assistance program. Significantly lower intakes of six nutrients were found among the food insufficient compared to the food sufficient. The percentage of calories supplied by carbohydrate was significantly higher among the food insufficient compared to the food sufficient.

When comparing demographic characteristics of food insufficient adults based on participation in food/nutrition assistance programs, those not participating in any program had a significantly higher mean poverty index ratio than either of the other two groups. Those with more formal educational attainment were least likely to participate in food assistance programs. Female-headed households were most likely to participate in two food assistance programs and those in the over 60 age group were least likely to participate in food assistance programs. Significantly lower intakes of four nutrients were found among those participating in only one food assistance program compared to those not participating, but not between those participating in one program and two programs or between two programs and no program.

There was a significant positive relationship between food insufficiency and percentage of calories from carbohydrate. There was a significant negative relationship between food insufficiency and 10 nutrients. Food/nutrition assistance program participation was significantly positively associated with increased intakes of 2 nutrients.

The NHANES survey design is a complex probability sampling design that is intended to be nationally representative. The Southern sites chosen for inclusion in NHANES III were in Florida and Texas. Therefore, there are limitations associated with generalizing these findings to the Southern region. In addition, the nationally representative nature of NHANES III may attenuate regional differences in diets. Our findings of lower intakes of some nutrients among the food insufficient generally agree with the findings of others that have used national survey data to study food insufficiency and nutrient intakes, with the exception of a higher percentage of calories supplied by carbohydrate and a positive relationship between food insufficiency and percent calories from carbohydrate. Future research should investigate whether or not this higher percentage of calories from carbohydrate is influencing micronutrient intakes in the Southern population. In addition, the finding of a positive relationship between food program participation and sodium intake deserves further investigation due to the possible adverse effects of high sodium diets on health.

Since regression analysis indicated a significant negative association between food insufficiency and nutrient intakes as well as a positive association between food program participation and sodium intakes, it is important that food and nutrition assistance programs continue to promote access to affordable and nutritious food for low-income families. In addition, continued emphasis on nutrition education such as that provided by WIC and the Family Nutrition Program may help to improve food choices and therefore nutrient intakes of those who participate in them. This will be a fruitful area for behavioral, educational, and program evaluation research in the future.