Food Insufficiency and Physical and Mental Health in ... - SAGE Journals

#1587—Jnl of Health and Social Behavior—Vol. 45 No. 2—45204-siefert

Food Insufficiency and Physical and Mental Health in a Longitudinal Survey of Welfare Recipients* KRISTINE SIEFERT University of Michigan

COLLEEN M. HEFLIN University of Kentucky

MARY E. CORCORAN University of Michigan

DAVID R. WILLIAMS University of Michigan

Journal of Health and Social Behavior, 2004, Vol 45 (June): 171–186

Food insufficiency is a significant problem in the United States, and poor African American women with children are at especially high risk. An inadequate household food supply can potentially affect the well-being of household members, but it is difficult to distinguish the effects of food insufficiency from risk factors for poor health that are also common among the food insufficient, such as poverty. We examined food insufficiency and physical and mental health among African American and white women (n = 676) who were welfare recipients in 1997. Controlling for common risk factors, women who reported food insufficiency in both 1997 and 1998 were more likely to report fair or poor health at the later date. Food insufficiency in 1998 was significantly associated with meeting the diagnostic screening criteria for recent major depression. Food insufficiency at both times and in 1998 only was related to women’s sense of mastery. These findings add to growing evidence that household food insufficiency is associated with poor physical and mental health. Despite the strength of the economy in recent years, high rates of food insecurity, food insufficiency, and hunger1 are significant problems in the United States (Alaimo et al.1998; Andrews et al. 2000). It is currently estimated that 31 million people live in food insecure * Support for this study was provided the Food Assistance and Nutrition Research Program of the United States Department of Agriculture’s Economic Research Service; the Institute for Research on Poverty Small Grants Program, University of Wisconsin; National Institute of Mental Health Grant No. R24-MH51361; and the Charles Stewart Mott and Joyce Foundations. Please address all correspondence to Kristine Siefert, NIMH Research Center on Poverty, Risk, and Mental Health, The University of Michigan, 1080 South University, Ann Arbor, MI 48109-1106; email: [email protected].

households, meaning that at some time during the previous year, they were unable to acquire or were uncertain of having enough food to meet basic needs due to inadequate household resources (Andrews et al. 2000). Food insecurity varies considerably by race/ethnicity: While only 7 percent of non-Hispanic white households in the United States reported food insecurity in 1999, 20.8 percent of Hispanic households and 21.2 percent of non-Hispanic black households were food insecure. Food insecurity was also higher among those in households with incomes below the poverty line, and among single mothers, who had rates of 36.7 percent and 29.7 percent, respectively. Rates of food insufficiency also vary considerably by race, gender, and socioeconomic status: Analyzing data from the Third National 171

#1587—Jnl of Health and Social Behavior—Vol. 45 No. 2—45204-siefert 172

JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

Health and Nutrition Examination Survey (NHANES III), Alaimo et al. (1998) found rates of food insufficiency to be 11.8 percent among low-income whites, 13.5 percent among low income non-Hispanic blacks, and 24.8 percent among low-income Mexican Americans; low-middle-income single femaleheaded families with children were 5.5 times more likely than other family types to be food insufficient. More recently, the United States Department of Agriculture found that, while the overall prevalence of food insecurity with hunger in the United States was 3.0 percent, it was 6.4 percent among black households, 5.5 percent among Hispanic households, 8.1 percent among families headed by a single woman, and 12.2 percent among households below the poverty line (Andrews et al. 2000). Policy changes in programs aimed at alleviating food insufficiency have resulted in increased concern regarding the causes and consequences of food insufficiency. For example, the enactment of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA) ended the federal guarantee of income support to low-income families with children. The legislation imposed a five-year lifetime limit on benefits, and recipients of Temporary Assistance to Needy Families (TANF) now must meet work requirements to receive benefits. The maximum food stamp benefit was reduced, and TANF recipients who leave welfare may stop receiving food stamp benefits altogether (Ziliak, Gunderson, and Figlio 2000; Mills et al. 2001; Quint et al. 2001). These changes have raised concerns about increased food insecurity and hunger and their effects on the well-being of current and recent welfare recipients (Borjas 2001). Declines in food insufficiency have been much lower than declines in welfare caseloads or receipt of food stamps. The number of food stamp recipients fell from a historic high of 27.5 million in 1994 to 17.4 million by the end of 1999 (Ziliak et al 2000); welfare caseloads fell by 50 percent over this same time period (U.S. Council of Economic Advisors 1999). In contrast, from 1995 to 1999 food insufficiency declined by only 16 percent (Andrews et al. 2000). Despite its potential impact on health and well-being, surprisingly little research has been done on the relationship between household food insufficiency and the physical and mental health of household members. This is

of particular concern in light of recent studies that have found that an inadequate household food supply is associated with lower energy and nutrient intakes among women. Rose and Oliveira (1997) analyzed the diets of 3,774 adult women who participated in the 1989–1991 Continuing Survey of Food Intake by Individuals and found that for adult women household food insufficiency was significantly associated with low energy and nutrient intakes. Although this finding was based on a single 24 hour recall and cannot be used to infer dietary inadequacy, the study did document a significant relationship between relative food deprivation and decreases in nutrient intake. The authors note that, given the limitations of nutritional epidemiology at this time, it may be difficult to ascertain the health impact of shortfalls in critical nutrients (Rose and Oliveira 1997). Their finding is important, however, as it documents that self-reported hunger measures are valid indicators of increased risk of dietary inadequacy (Sidel 1997). In a study of 145 women in families receiving emergency food assistance in Toronto, Tarasuk and Beaton (1999) found that those who reported hunger in their household during the past month also reported systematically lower intakes of energy and nutrients, despite controlling for a broad range of potential economic, sociocultural, and behavioral influences on diet. Although these findings must be interpreted with caution due to the small sample size and other methodological limitations, the observed association between systematically low reported nutrient intakes and more severe food insecurity (food insecurity with moderate or severe hunger), independent of other potential influences on diet, does imply that women in households experiencing severe food insecurity may be at particular risk of nutrient deficiencies. The authors also note that, although the immediate impact of low nutrient intakes may be minimal, the long term effect of chronically low intakes is of concern, as low intakes of some nutrients are associated with chronic disease. This suggests that it is important to distinguish between temporary and persistent or recurrent food insufficiency when examining the relationships between food insufficiency and health. More recently, Dixon, Winkelby, and Radimer (2001) found that adults in food insufficient households were more likely to

#1587—Jnl of Health and Social Behavior—Vol. 45 No. 2—45204-siefert FOOD INSUFFICIENCY AND PHYSICAL AND MENTAL HEALTH have diets that could compromise their health than adults from food insufficient families, and in addition had lower serum concentrations of critical nutrients. Low intakes of such nutrients increase the risk of a number of major chronic diseases (Dixon et al. 2001). STUDY DESIGN AND HYPOTHESES Although food insecurity and hunger can potentially affect mental well-being and overall quality of life, as well as health outcomes, an important challenge facing research on nutrition and health is to distinguish the health consequences of food insecurity from those of its common risk factors, such as poverty and low socioeconomic status (Campbell 1991; Olson 1999). A recent study by Siefert et al. (2001) analyzed the relationship between household food insufficiency and indicators of physical and mental health status among 733 African American and white women who were current or recent welfare recipients. This study found that household food insufficiency was a significant predictor of fair or poor self-rated health, limitations in physical functioning, and meeting the DSM-III-R diagnostic criteria for major depression, while controlling for other factors known to be associated with lowincome women’s health and mental health. Although race was not a significant predictor of physical health status in this study, consistent with other research (Somervell et al. 1989; Williams et al. 1992; Kessler et al. 1994), being African American did decrease the odds of depression, even in this economically disadvantaged sample. This study was limited by its cross-sectional design, however. The present study investigates the physical and mental health effects of food insufficiency by using two waves of the same data set as the cross-sectional study of welfare recipients described above and extends the earlier study in three respects. First, longitudinal data permit us to construct measures of persistent or recurrent food insufficiency. Second, longitudinal data enable us to estimate effects of food insufficiency on health status at wave 2, controlling for both initial health status at wave 1 and personal characteristics and common risk factors. Third, the cross-sectional study examined only physical and mental health problems. In this study, we also examine the effects of household food insufficiency on a positive

173

measure of psychological functioning: mastery, or the degree to which respondents feel that in they are in control of their lives. It is likely that food insufficiency could adversely affect an individual’s sense of mastery. Although this issue has received little empirical attention in the past, there are several mechanisms by which this might occur. First, food insufficiency may indirectly affect mental health because of its adverse impact on physical health. Poor physical health is a risk factor for poor mental health such that poor physical health status may mediate the relationship between food insufficiency and mental health. Second, food insufficiency may be subjectively experienced as a stressor, and its persistence may both increase feelings of selfblame and heighten the individual’s perception that s/he is not efficacious. Prior research indicates that an individual’s sense of mastery is largely a consequence of experiencing oneself as efficacious (Gecas and Schwalbe 1983), and exposure to stressful life experiences can erode feelings of mastery (Krause and Tran 1989). Third, food insufficiency could also have a negative impact on mastery through a direct effect of nutritional deficiency on psychological status and behavior. Prior research documents alterations in behavior and mental performance even in the early stages of nutrient deficiency. For example, an experimental study of 1,081 healthy young men found that reduced vitamin intake over a two month period was associated with changes in psychological dispositions and functioning (Heseker et al. 1992). Specifically, inadequate vitamin intake was associated with increased irritability, nervousness, depression, and feelings of fear, as well as decreased well-being, memory, and reaction performance. Instructively, several of these effects were reversed by vitamin supplementation. Our conceptual framework draws on epidemiological theories of the social production of health and disease (Kreiger and Zierler 1996; Link and Phelan 1995; Williams 1997), which posit that individuals’ relative social and economic positioning determines their exposures to health risks and shapes their health behavior, and that the distribution of disease in populations reflects this positioning. Welfare recipients are disproportionately African American, they are predominantly female, and by definition they are poor. We hypothesized that household food insufficiency would be



associated with worse physical and mental health in this population, while controlling for baseline health status, as well as for poverty and other factors known to affect physical and mental health. We further hypothesized that persistent or recurrent food insufficiency would be associated with worse physical and mental health status, but based on the crosssectional findings described above, we also hypothesized that being African American would decrease the risk of depression in this sample. METHODS Study Design and Sample We analyzed data from the second wave of the Women’s Employment Study, a panel survey of barriers to employment among 753 mothers who were receiving cash assistance in an urban Michigan county in February 1997. Trained staff of the Survey Research Center of the Institute for Social Research of the University of Michigan conducted face-toface, in-home, structured interviews lasting approximately one hour between August and December of 1997 and again in 1998. Women were eligible if they resided in the study county, received cash assistance in February of 1997, were single and a U.S. citizen between the ages of 18 and 54, and claimed a racial identity of African American or white (there were too few other minority residents of this county to conduct reliable analyses). A simple random sampling scheme was used. Cases were systematically selected with equal probability from an ordered list of eligible single mothers. To derive a representative sample of the metropolitan area and the study population, cases were proportionately selected by zip code, race (African American or nonHispanic white), and age. The response rate was 86.2 at wave one and 92 percent at the second wave. Our sample consists of 676 women who were interviewed in both waves and have no missing data on any of the variables used in the analyses. In analyses not shown here (available from the authors upon request), we compared characteristics of the study sample with those of respondents who were not reinterviewed in 1998. Surveyed respondents did not differ significantly from attriters on characteristics such as race, food insufficiency, and the

four outcomes examined here. Given this and the low attrition rate between waves 1 and 2, the risk of attrition bias was considered low and the data were not weighted. Variables Assessed and Definitions Food insufficiency was defined narrowly as restricted household food stores or too little food intake among either adults or children in the household (Scott and Wehler 1998) and was operationalized using the question, “Which of the following describes the amount of food your household has to eat—enough to eat, sometimes not enough to eat, or often not enough to eat?” This single-item measure is widely accepted as a valid measure of food insufficiency (Alaimo et al. 1998; Rose 1999). Studies by Christofar and Basiotis (1992) and Rose and Oliveira (1997) analyzed 24-hour dietary recall data and found that individuals who responded “sometimes” or “often not enough to eat” were significantly less likely to consume the percentage of recommended dietary allowances for energy, protein, vitamins, and minerals and consumed lower gram amounts of milk and milk products, vegetables, and fruits. Dixon et al. (2001) analyzed both dietary intakes and serum concentrations of nutrients between adults from food insufficient and food sufficient families and found that adults in food insufficient families were more likely to have lower intakes of critical nutrients and foods; they also found that adults in food insufficient families were more likely to have very low serum concentrations of critical nutrients. Following the convention that has been adopted in related research (Christofar and Basiotis 1992; Rose and Oliveira 1997; Alaimo et al. 1998), we coded respondents as food insufficient who answered “sometimes” or “often.” Because we have panel data, we created four dummy variables to control for the presence or absence of food insufficiency over time: Persistent or recurrent food insufficiency indicates those food insufficient at both waves 1 and 2; wave 1 food insufficient indicates those food insufficient at wave 1, but not at wave 2; wave 2 food insufficient indicates those food sufficient at wave 1, but insufficient at wave 2; and never food insufficient, the omitted group in our regression analyses, indicates those food sufficient at both waves.

#1587—Jnl of Health and Social Behavior—Vol. 45 No. 2—45204-siefert FOOD INSUFFICIENCY AND PHYSICAL AND MENTAL HEALTH Our dependent variables included both positive and negative indicators of physical and mental health and well-being: global self-rated health, limitations in physical functioning, major depression, and sense of mastery. Single-item measures of self-rated health have been shown to be reliable and valid predictors of mortality and morbidity when controlling for other health status indicators (Idler and Benyamini 1997). We created a dummy variable coded as 1 for women who described their overall physical health as “fair” or “poor,” as opposed to “excellent,” “very good,” or “good.” Physical functioning limitations were assessed using items from the SF-36 Physical Functioning subscale (Ware and Sherbourne 1992). Respondents who scored in the lowest age-specfic quartile (based on general population norms) were defined as having physical functioning limitations. The measurement of major depressive disorder, which is common among women and associated with significant impairment in social and occupational functioning, was operationalized in screening versions of the World Health Organization’s Composite International Diagnostic Interview (WHO 1990; Kessler et al.–1998). The Composite International Diagnostic Interview (CIDI) is a structured interview schedule designed to be used by trained interviewers who are nonclinicians to assess the prevalence of specific psychiatric disorders. World Health Organization field trials and other methodological studies (Blazer et al. 1994; Wittchen 1994) have documented acceptable test-retest reliability and clinical validity of CIDI diagnoses. We measured mastery, which refers to the degree to which individuals perceive themselves to be in control of their own lives, by using the seven-item Pearlin Mastery Scale (Pearlin et al.1981), a widely used measure of this construct. We defined high mastery as the top quartile of the wave 1 distribution. Our independent variables included six sociodemographic and personal characteristics known to be associated with health and mental health among low income women: race, age, number of children in the household, education level, single parent status, and length of time on welfare prior to participating in the survey. Race is a dummy variable coded 1 for African American. Anticipating non-linearities in the effect of age on women’s physical and mental

175

health, we broke age into three categories: age 18–24 as the omitted reference group, age 25–34, and age 35 and older. The number of children in the household was controlled as a dummy variable indicating those families that have more than the median of two children in the household. Education level was entered as a dummy variable coded 1 for those with less than a high school diploma or General Educational Development certificate. A woman was coded as married if she is married or living with her partner. Welfare history was measured by a dummy variable that equals 1 if the women had received welfare for more than two years prior to wave 1. With the exception of the welfare variable, all background characteristics were measured at wave 1. We also controlled for a broad array of social and environmental risk factors related to race, gender, and social class and known to be associated with increased risk of poor health among women and also known to be or likely to be associated with food insufficiency. These factors include poverty, not being employed, and poverty-related stressful life circumstances, such as homelessness or utility shutoffs (Bartley 1994; Bassuk et al. 1996; Kendall, Olson, and Frongillo 1996; Alaimo et al. 1998; Corcoran, Heflin, and Siefert 1999; Siefert et al. 2001). In addition, we controlled for domestic violence, which has been shown to adversely affect physical and mental health among women and has also been associated with food insufficiency (Roberts et al. 1998; Corcoran et al. 1999; Siefert et al. 2001); and for self-reported experiences of discrimination based on race or gender. Perceived discrimination has been associated with poor physical and mental health (Williams et al. 1997; Kessler et al. 1999), and, as noted, African American women are disproportionately represented among the food insufficient. All risk factors were measured at wave 1. Poverty status was measured using the official ratio of total family income from all sources and all household members to the federal poverty line for a given family size (U.S. Bureau of the Census 1998). Any woman whose income, plus the estimated value of the Earned Income Tax Credit and including the value of Food Stamps, was above the annual poverty line was considered not poor. Not employed was defined as a negative response to the question, “Are you currently working for pay?”



Stressful life circumstances were defined as four or more positive responses to an 11-item measure which we adapted for this study from the Difficult Life Circumstances scale, which was developed by Booth et al. (1989) and adapted for use in the New Chance study of young mothers in poverty by Quint, Bos, and Polit (1997). Eight items were drawn from this scale, which measures ongoing or habitual stressors that are often a feature of living in poor communities. An example of an item is, “Have you had trouble finding a place to live?” Three additional questions obtained information about various material hardships experienced over the past year, including utility shutoffs, eviction, and homelessness. Domestic violence was measured using items drawn from the Conflict Tactics Scale (Straus and Gelles 1986), a widely-used measure of family violence. In the present study, respondents were defined as having experienced domestic violence if they reported experiencing any of six indicators of severe physical violence by a partner within the past year (hit with a fist, hit with an object, beaten, choked, threatened or assaulted with a weapon, or forced into sexual activity). Questions about perceived discrimination based on race or gender were adapted from items used in surveys by Bobo (1995) and Williams et al. (1997). Discrimination based on race was defined as a positive response to any of a series of five questions which asked about unfair treatment while seeking employment or in the workplace, including whether respondents thought they had ever been refused a job, fired, or not promoted because of their race. An example of an item is, “Did your supervisor or boss ever use racial slurs?” Discrimination based on gender was defined a positive response to a series of six questions which asked about unfair treatment while seeking employment or in the workplace, including whether respondents thought they had ever been refused a job, fired, or not promoted because of their sex, whether they had been sexually harassed, and if their supervisor made insulting remarks about women. Data Analysis We estimated three nested logistic regression models to examine the independent effects of three sets of factors on our four

dependent variables. In model 1, we began by showing the bivariate relationships between food insufficiency and each of the dependent variables. This represents the uncontrolled effect of food insufficiency shown in the bivariate tables. In model 2, we controlled for the background demographic characteristics and social and environmental risk factors described above. In model 3, we controlled for initial health status. Including the wave 1 version of the dependent variable changes the interpretation of the odds ratios in model 3 to, for example, the effect of food insufficiency on the change in health status (Plewis 1985). Using SPSS Version 9.01 software, we computed odds ratios and 95 percent confidence intervals for unit changes in each factor. RESULTS Food Insufficiency and Sample Characteristics Over one-third of the respondents had experienced food insufficiency at one or both of the interviews: 11.8 percent were food insufficient at both waves, 12.7 percent were food insufficient only at wave 1, and 10.2 percent were food insufficient only at wave 2. The first column of numbers in Table 1 describes demographic characteristics of the sample, the prevalence of social and environmental risk factors, and physical and mental health status at wave 2. Fifty-five percent of the respondents were African American, 30.9 percent were age 35 years or older, 37.7 percent had three or more children, and 29.0 percent had not completed high school. About two-thirds were unmarried and without a partner, and 80.3 percent had been on welfare for more than two years prior to the wave 1 interview. Risk factors for poor health and mental health were highly prevalent. Half the sample was poor in the month prior to the survey, and 34.9 percent were not employed at the time of the wave 1 interview. Over 18 percent of the sample had experienced four or more stressful life circumstances, and 15.5 percent had experienced severe abuse in the last year. Roughly one in seven respondents reported experiencing racial discrimination, and the same proportion reported experiencing sex discrimination. Twenty-eight percent of women in the sample reported being in fair or poor health at wave 2;

53.6 52.2 23.2 46.4 39.1 29.0 82.6 37.7 47.8 29.0 15.9 18.8 23.2 33.3 53.6 30.4 17.4

55.8 41.9 32.6 38.4 34.9 29.1 80.2 52.3 32.6 18.6 22.1 17.4 14.0 25.6 45.3 11.6 29.1

Prevalence among Prevalence among food insufficient at food insufficient at wave 1 only (12.7%) wave 2 only (10.2%)

Variable Background Characteristics 54.9 (.50) African American 46.0 (.50) Age 25–34 30.9 (.46) Age 35 and older 37.7 (.49) Number of children (if 3 or more in home) 29.0 (.45) Less than high school education 33.0 (.47) Lives with husband or partner 80.3 (.40) On welfare more than 2 years at wave 1 Social and Environmental Risk Factors 49.6 (.50) Poverty status (1 = not poor) 34.9 (.48) Not employed at time of survey 18.6 (.39) Stressful life circumstances 15.5 (.36) Domestic violence 16.6 (.37) Sex discrimination 16.3 (.37) Race discrimination Physical and Mental Health at W2 28.4 (.45) Fair or poor self-rated health 45.1 (.50) Limitations in physical functioning 16.4 (.37) Major depressive disorder 33.3 (.47) High sense of mastery Note: P-values from chi-square test for independence between rows and columns.

Prevalence in sample (SD)

47.5 62.5 28.8 17.5

32.5 47.5 32.5 28.8 16.3 11.3

54.4 38.8 47.5 42.5 40.0 31.3 88.8

Prevalence among food insufficient at both waves (11.8%)

TABLE 1. Sample Characteristics, Risk Factors, and Physical and Mental Health Status by Food Insufficiency (N = 676)

24.7 40.6 12.9 39.5

54.0 31.1 14.5 11.8 16.1 16.6

54.9 47.2 28.8 35.4 24.3 35.6 78.5

Prevalence among food sufficient (65.2%)

.000 .001 .000 .000

.001 .003 .000 .000 .944 .234

.954 .310 .004 .258 .002 .486 .188

p-values

#1587—Jnl of Health and Social Behavior—Vol. 45 No. 2—45204-siefert

FOOD INSUFFICIENCY AND PHYSICAL AND MENTAL HEALTH 177



45.1 percent reported limitations in physical functioning, and 16.4 percent met the diagnostic screening criteria for recent major depressive episode. A positive finding was that 33.3 percent reported a high sense of mastery. The last four columns of numbers in Table 1 report the distribution of background characteristics, risk factors, and health status indicators by food insufficiency status across waves 1 and 2. These relationships were significant for age, education, poverty status, not being employed, stressful life circumstances, domestic violence, fair or poor self-rated health, physical functioning limitations, major depressive disorder, and high sense of mastery. Women who did not report food insufficiency at either wave were more likely than women who reported food insufficiency at both waves and women who reported food insufficiency only at wave 2 to: be younger, to have a high school diploma, to not be poor, to be employed, to not report domestic violence, to not report four or more stressful life circumstances, to not report limitations in physical functioning, to not rate their overall health as fair or poor, to not meet the criteria for major depression, and to have higher mastery scores. For the most part, the background characteristics, risk factors, and health and mental health status of women who had reported food insufficiency only at wave 1 resembled those of women who had never reported food insufficiency. Table 2 presents the distribution of the four outcomes variables over the two-wave observation period. Each of the outcomes examined showed a high degree of change, indicating that physical and mental health are highly dynamic outcomes. For example, while 17.6 percent of the sample met the diagnostic criteria for major depression at wave 1 but then failed to meet the criteria at wave 2, 7.7 percent of women who had not met the diagnostic criteria at wave 1 met the criteria at wave 2. Sixty-six percent of women did not meet the criteria for depression at either wave, and 8.7 met the criteria at both waves. The pattern is

roughly similar for mastery and self-rated health. Limitation in physical functioning deserves special attention because a high proportion (31.5 %) of women reported higher levels of such limitations at both waves. Tables 3 through 6 present the results from the nested logistic regression analyses. Table 3 presents results predicting fair or poor selfrated health. The results in Table 3 suggest that the timing of food insufficiency is important. Food insufficiency at both waves was significantly and negatively associated with self-ratings of health at wave 2 when nothing was controlled (model 1), when individual characteristics and risk factors were controlled (model 2), and when self-ratings of health at wave 1 were controlled (model 3). The size of the association between persistent or recurrent food insufficiency and self-rated health at wave 2 dropped substantially when background characteristics and risk factors were controlled. Although being food insufficient only at wave 2 or only at wave 1 was never significantly related to self-rated health at wave 2, the pattern of the odds ratios is interesting and suggests that the timing of food insufficiency matters: Being food insufficient at wave 1 only was not associated with self-rated health at wave 2, but being food insufficient only at wave 2 was somewhat associated with health at wave 2, and being food insufficient at both waves was more associated with health at wave 2. As expected, some of the background and risk factors also affected self-ratings of health. Older women, women not employed, and women who reported four or more stressful life circumstances had significantly higher odds of reporting fair or poor health at wave 2 than did other women; women who were not poor and those with three or more children had lower odds of fair or poor health. However, the effects of the number of children and being employed were no longer statistically significant once self-rated health at wave 1 was controlled (model 3). Not surprisingly, self-rated health at wave 1 was strongly and significant-

TABLE 2. Physical and Mental Health Status over Time

Fair/Poor Self-Rated Health Physical Limitations Depression High Mastery

Condition Present at Wave 1 Only 09.6 13.6 17.6 11.8

Condition Present at Wave 2 Only 12.0 15.2 07.7 15.7

Condition Present at Wave 1 & Wave 2 16.4 31.5 08.7 17.6

Condition Never Present 62.0 39.6 66.0 54.9

#1587—Jnl of Health and Social Behavior—Vol. 45 No. 2—45204-siefert FOOD INSUFFICIENCY AND PHYSICAL AND MENTAL HEALTH

179

TABLE 3. Food Insufficiency and Fair or Poor Self-Rated Health (N = 676) Model 1 OR

95% CI

Food Insufficiency Food insufficient W1 only 1.05 .62–1.78 Food insufficient W2 only 1.52 .88–2.63 Food insufficient both waves 2.76*** 1.69–4.49 Wave 1 Background Characteristics African American Age 25–34 Age 35 and older Number of children (if 3 or more in home) Less than high school education Lives with husband or partner On welfare >2 years at wave 1 Wave 1 Social and Environmental Risk Factors Poverty status (1 = not poor) Not employed at time of survey Stressful life circumstances Domestic violence Sex discrimination Race discrimination Lagged Fair/Poor Health –2 Log likelihood 789.6 df (n – k) 673 † p < .10; * p < .05; ** p < .01; *** p < .001

ly associated with self-rated health at wave 2. Interestingly, adding health status at wave 1 into the model also caused the coefficient on racial discrimination to reach statistical significance, suggesting that self-rated health at

Model 2

Model 3

OR

95% CI

OR

95% CI

1.00 1.24 1.93*

.57–1.74 .69–2.22 1.13–3.28

.77 1.24 1.70†

.42–1.43 .65–2.35 .93–3.09

.78 .53–1.15 1.28 .73–2.23 2.30** 1.29–4.12 .68† .46–1.01 1.30 .87–1.94 1.12 .76–1.17 1.15 .66–2.02 .57** .38–0.85 1.44† .97–2.14 1.87** 1.18–2.95 1.08 .66–1.77 .97 .57–1.66 1.36 .80–2.32 738.6 660

.80 .52–1.21 1.38 .75–2.55 1.96* 1.02–3.76 .71 .46–1.09 1.13 .73–1.77 1.12 .73–1.73 .93 .50–1.72 .55** .35–0.86 1.33 .86–2.07 1.61† .97–2.65 1.05 .61–1.80 .89 .50–1.58 2.09* 1.18–3.71 8.32***5.46–12.67 632.1 659

wave 1 is correlated with perceived employment discrimination on the basis of race. Table 4 presents the results from the logistic regressions predicting physical functioning at wave 2. Food insufficiency at both waves sig-

TABLE 4. Food Insufficiency and Limitations in Physical Functioning (N = 676) Model 1 OR 95% CI Food Insufficiency Food insufficient W1 only 1.21 .76–1.93 Food insufficient W2 only 1.69* 1.02–2.82 Food insufficient both waves 2.44*** 1.49–3.99 Wave 1 Background Characteristics African American Age 25–34 Age 35 and older Number of children (if 3 or more in home) Less than high school education Lives with husband or partner On welfare >2 years at wave 1 Wave 1 Social and Environmental Risk Factors Poverty status (1 = not poor) Not employed at time of survey Stressful life circumstances Domestic violence Sex discrimination Race discrimination Lagged Physical Functioning –2 Log likelihood 915.3 df (n – k) 673 † p < .10; * p < .05; ** p < .01; *** p < .001

Model 2

Model 3

OR

95% CI

OR

95% CI

1.14 1.45 1.90*

.60–1.65 .79–2.30 1.20–3.42

.90 1.24 1.50

.53–1.52 .68–2.26 .85–2.64

.99 1.01 1.66† .94 1.49* 1.00 .74

.70–1.41 .63–1.63 .99–2.78 .66–1.33 1.03–2.16 .70–1.42 .46–1.20

.99 1.18 1.55 .93 1.35 .98 .75

.68–1.44 .71–1.98 .89–2.70 .64–1.35 .90–2.01 .67–1.44 .45–1.26

.93 .65–1.33 2.55*** 1.77–3.69 1.20 .77–1.86 .98 .62–1.57 .91 .56–1.46 .93 .57–1.51 859.2 660

.91 .62–1.34 2.34*** 1.58–3.48 1.04 .65–1.67 .85 .51–1.41 1.02 .61–1.71 1.00 .60–1.69 5.16*** 3.63–7.35 770.4 659



nificantly predicted worse physical functioning at wave 2 in the bivariate case (model 1) and when individual characteristics and social-environmental risk factors were controlled (model 2), but became insignificant once we control for physical functioning at wave 1 (model 3). Food insufficiency at wave 2 only was associated with limitations in physical functioning in the bivariate model, but declined in statistical significance once we added in the model 2 controls. Once again, food insufficiency only at wave 1 was not significantly related to physical functioning, even in the bivariate case. And when we further controlled for physical functioning at wave 1 both age and education became insignificant. Not being employed at the time of the wave 1 interview was associated with physical functioning in both models 2 and 3. As anticipated, physical functioning at wave 1 was a highly significant predictor of physical functioning at wave 2. Results for major depression at wave 2 are shown in Table 5. Food insufficiency at both waves and food insufficiency only at wave 2 were significantly associated with meeting the diagnostic screening criteria for major depression at wave 2 (model 1). When background characteristics and social and environmental risk factors were controlled, these associations became smaller in size but remained signifi-

cant. When initial depression status was controlled, however, the coefficient on persistent food insufficiency became insignificant, while the coefficient on food insufficiency only at wave 2 remained significant (model 3). Food insufficiency only at wave 1 was not significantly correlated with depression in wave 2 in any of the three models. Exposure to stressful life circumstances, domestic violence, and not being employed at the time of the wave 2 interview also significantly increased the likelihood of depression, although only the effect of stressful life circumstances remained significant after controlling for depression status at wave 1 (model 3). As expected, depression status at wave 1 strongly predicted depression status at wave 2. The results for sense of mastery, shown in Table 6, suggest a strong role for food insufficiency. Both persistent or recurrent food insufficiency and food insufficiency only at wave 2 were consistently related to a high sense of mastery in models 1 through 3. That is, women who were food insufficient only at wave 2 and women who were food insufficient at waves 1 and 2 were much less likely to have high mastery scores at wave 2, even when background characteristics, social-environmental risk factors, and mastery at wave 1 were controlled. Food insufficiency only at wave 1 was insignificant in each of the models. Older age and

TABLE 5. Food Insufficiency and Recent Major Depression (N = 676) Model 1 OR 95% CI Food Insufficiency Food insufficient W1 only .89 .43–1.81 Food insufficient W2 only 2.95*** 1.80–5.77 Food insufficient both waves 2.72*** 1.56–4.75 Wave 1 Background Characteristics African American Age 25–34 Age 35 and older Number of children (if 3 or more in home) Less than high school education Lives with husband or partner On welfare >2 years at wave 1 Wave 1 Social and Environmental Risk Factors Poverty status (1 = not poor) Not employed at time of survey Stressful life circumstances Domestic violence Sex discrimination Race discrimination Lagged Major Depression –2 Log likelihood 582.1 df (n – k) 673 † p < .10; * p < .05; ** p < .01; *** p < .001

Model 2 OR

95% CI

Model 3 OR

95% CI

.74 .35–1.57 2.50*** 1.34–4.68 1.88* 1.02–3.49

.62 .29–1.33 2.52** 1.32–4.80 1.39 .73–2.65

.67† 1.31 1.44 1.07 .95 .96 1.27

.75 1.20 1.33 1.04 .99 .98 1.28

.42–1.07 .66–2.62 .69–3.02 .67–1.71 .58–1.58 .60–1.55 .63–2.56

1.35 .81–2.23 1.71* 1.04–2.80 2.99*** 1.80–4.95 1.71† .98–2.97 1.37 .74–2.54 .60 .30–1.23 541.8 660

.46–1.22 .59–2.44 .33–2.83 .65–1.68 .59–1.66 .60–1.60 .62–2.61

1.38 .83–1.66 1.52 .91–2.52 2.60*** 1.54–4.38 1.50 .85–2.67 1.30 .70–2.42 .62 .30–1.26 3.21*** 2.01–5.13 518.4 659

#1587—Jnl of Health and Social Behavior—Vol. 45 No. 2—45204-siefert FOOD INSUFFICIENCY AND PHYSICAL AND MENTAL HEALTH

181

TABLE 6. Food Insufficiency and Self-Reported Mastery (N = 676) Model 1 OR Food Insufficiency Food insufficient W1 only Food insufficient W2 only Food insufficient both waves Wave 1 Background Characteristics African American Age 25–34 Age 35 and older Number of children (if 3 or more in home) Less than high school education Lives with husband or partner On welfare >2 years at wave 1 Wave 1 Social and Environmental Risk Factors Poverty status (1 = not poor) Not employed at time of survey Stressful life circumstances Domestic violence Sex discrimination Race discrimination Lagged High Mastery –2 Log likelihood df (n – k) † p < .10; * p < .05; ** p < .01; *** p < .001

95% CI

.63 .38–1.04 .32*** .17–.062 .33*** .18–0.60

833.2 673

lower education were associated with having a lower sense of mastery at wave 2; having three or more children was associated with a higher sense of mastery. Of the social and environmental risk factors, domestic violence and not being employed at the wave 1 interview were associated with decreased odds of having a sense of control over one’s life. After controlling for mastery at wave 1, however, only domestic violence remained marginally significant. Finally, a high sense of mastery at wave 1 was strongly related to high mastery at wave 2. In analyses not shown but available from the authors, race was interacted with the three food insufficiency variables. For each of the outcomes examined, none of the race interactions were statistically significant. Based on our cross-sectional findings, we had hypothesized that being African American would reduce the risk of depression, a finding that speaks to the importance of longitudinal analysis. We also tested whether physical health might mediate the relationship between food insufficiency and mental health. We entered each measure of physical health separately into model 3 for the analyses predicting depression and mastery. In each case, physical health had no effect on the

Model 2

Model 3

OR

95% CI

.68 .35** .44*

.40–1.15 .18–0.68 .23–0.83

.77 .45–1.34 .29*** .14–0.58 .45* .23–0.88

1.19 .83–1.73 .63† .39–1.03 .42** .24–0.72 1.49* 1.03–2.14 .58* .39–0.88 1.00 .69–1.45 1.23 .74–2.05

1.17 .79–1.74 .62† .37–1.03 .40** .22–0.71 1.55* 1.05–2.28 .66† .42–1.02 1.00 .67–1.49 1.33 .78–2.27

1.25 .68† .85 .62† 1.08 1.07

1.25 .80 .92 .55* 1.04 .93

793.1 660

.86–1.83 .45–1.02 .53–1.37 .36–1.06 .65–1.77 .65–1.76

OR

95% CI

.84–1.88 .52–1.23 .55–1.52 .31–0.98 .62–1.77 .54–1.60

5.07*** 3.46–7.42 719.5 659

pattern of association between food insufficiency and the mental health outcomes. DISCUSSION Although the limitations of our measures and the constraints imposed by self-reported data must be considered in interpreting our results, the findings reported here suggest that household food insufficiency matters for the physical and mental health of these lowincome African American and white women. The finding that persistent or recurrent food insufficiency is a significant and independent predictor of self-rated health in this sample has potentially serious implications. Self-rated health is a widely used and well-validated predictor of subsequent mortality and morbidity: In their review of 27 prospective studies of representative community-based samples selected for their excellent research designs, Idler and Benyamini (1997) found that global self-rated health was an independent predictor of mortality while controlling for a broad array of health status indicators and correlates, and they concluded that self-ratings represent an “irreplaceable dimension of health status” (p. 34). Selfrated health has also been found to predict



functional decline and chronic disease (Idler and Kasl 1995). As noted, Rose and Oliveira (1997) found that household food insufficiency was significantly associated with decreases in nutrient intakes, and Tarasuk and Beaton (1999) found that women experiencing food insecurity with moderate or severe hunger reported low nutrient intakes. If household food insufficiency is in fact a valid indicator of increased risk of dietary inadequacy, as Rose and Oliveira’s research suggests, our findings are consistent with the hypothesis that persistent or recurrent household food insufficiency damages women’s health through low nutrient intakes. Our results with respect to food insufficiency and meeting the diagnostic screening criteria for major depression are particularly interesting in light of current research on the etiology and course of this disorder. In addition to the possibility that household food insufficiency could directly affect depression through nutritional deprivation, it is possible that anxiety about not having enough food in the household is a stressful experience that could dysregulate the hypothalamic-pituitary-adrenocortical axis (McEwen 1998). The HPA axis is highly responsive to stress, and social and environmental factors are known to confer increased risk (Bruce, Takeuchi, and Leaf 1991; Williams et al. 1992; Blazer et al. 1994; Siefert et al. 2000). Approximately half of those who experience a major depressive disorder recover within a year, but the illness becomes recurrent in 50 to 70 percent of those who experience one episode, and multiple episodes increase both risk of recurrence and severity (Wells et al.1996). We found that those women who became food insufficient between waves 1 and 2 were more likely to meet the diagnostic criteria for major depression at wave 2. If confirmed by further research, this suggests that preventing food insufficiency might lower the risk of major depression in this sample. Our findings with respect to respondents’ sense of mastery are also interesting, given the findings of previous research on this widelyused measure of psychological status. Both persistent or recurrent food insufficiency and food insufficiency only at wave 2 strongly predicted the extent to which women perceived a sense of control over their lives, even when we controlled for their sense of mastery at wave 1. Mastery captures beliefs about one’s ability to

control important outcomes in life. It is widely viewed as a fundamental characteristic of human beings that affects the ability of individuals to influence their environment (Rothbaum, Weisz, and Snyder 1982). High levels of mastery and control have been inversely related to socioeconomic status (Mirowsky and Ross 1986) and positively related to physical and mental health (Rodin 1986). As noted above, it is possible that food insufficiency may affect mental health through a direct effect of nutritional deprivation on psychological functioning. However, as Gecas and Schwalbe (1983) have observed, efficacy and mastery may largely be a product of experiencing oneself as efficacious, and if there is not sufficient food in the household, a respondent may blame herself, a process that could also contribute to depression. In summary, although severe malnutrition is rare in the United States, as Rose and Oliveira (1997) note, relative food deprivation clearly exists, and although we cannot infer causality given our observational study, our findings suggest that the association between household food insufficiency and physical and mental health has potentially serious implications. Importantly, our study suggests that lacking food even at a level that does not approach severe deprivation still has significant effects on physical and mental health. Our results also suggest that the effects of household food insufficiency on health and mental health may not be permanent if the food insufficiency is short-term. The physical and mental health outcomes of women who reported food insufficiency at wave 1 but not at wave 2 were not significantly different from those of women who never reported food insufficiency, once background characteristics and risk factors were controlled. As Gecas and Schwalbe (1983) point out, social-structural conditions can promote or constrain efficacious action. Our findings suggest that policies enabling women who head households that experience food insufficiency to obtain adequate food for their families in a timely manner might prevent or reverse the adverse effects of food insufficiency on their health and well-being. Our findings also highlight the need for more research to better understand the relationship between household food insufficiency and the nutritional status of household members, as well as the immediate and long-term consequences of even mild to moderate nutri-

#1587—Jnl of Health and Social Behavior—Vol. 45 No. 2—45204-siefert FOOD INSUFFICIENCY AND PHYSICAL AND MENTAL HEALTH tional deprivation on physical and mental health. As noted, although self-reported household food insufficiency does not imply severe malnutrition, it has been associated with lower serum concentrations of critical nutrients in adults, which elevates the risk of a number of major chronic conditions (Dixon et al. 2001). If household food insufficiency does in fact damage health through a direct effect of nutritional deficiency, it is encouraging to note that research suggests that policies to ensure adequate nutrition are likely to have an immediate impact in improving physical and mental functioning. One study found that nutritional supplementation over a two month period for men manifesting a mild to moderate vitamin deficiency led to improvements on several psychological factors including self-confidence, extraversion, concentration, and mood (Hesecker et al. 1992). Interestingly, a recent, comprehensive blueprint for reducing social inequalities in health, developed in response to a request from the British government, highlighted the importance of policies that seek to improve the nutritional status of vulnerable populations (Acheson 1998). This report calls for policies that seek to increase the availability, accessibility, and affordability of wholesome foods to disadvantaged populations. Finally, consistent with other research, we found that social and environmental risk factors related to the relative social and economic positioning of the women in our sample were significant predictors of self-reported physical and mental health. These findings imply that there are multiple risk factors associated with the social context that are linked to elevated rates of disease in this population, and they indicate the need for multifactorial and comprehensive population level interventions to shift the structural conditions that produce pathogenic exposures (Krieger and Zierler 1996). As Link and Phelan (1995) observe, ameliorative policies are unlikely to eliminate health disparities unless the fundamental causes of those disparities are addressed, and such policies are therefore necessary, but not sufficient. NOTE 1. Survey researchers use various terms to describe food scarcity and deprivation. Food insecurity” refers to the limited or

183

uncertain availability of food, while “food insufficiency” refers to restricted household food stores or insufficient food intake (Scott and Wehler 1998). “Hunger” refers to insufficient food intake. Food insecurity differs conceptually from food insufficiency and hunger, which can be considered roughly equivalent (Rose 1999). The distinction between food insufficiency and food insecurity can best be understood from a temporal frame of reference: Food insecurity can be experienced prior to the onset of food insufficiency, and it may or may not result in food insufficiency (Scott and Wehler 1998). Food insecurity with hunger, a measure recently adopted by the United States Department of Agriculture, is conceptually comparable to food insufficiency (Dixon, Winkleby, and Radimer 2001). REFERENCES Acheson, Donald. 1998. Independent Inquiry into Inequalities in Health: Report. London: Her Majesty’s Stationery Office. Alaimo, Katherine, Ronette R. Briefel, Edward A. Frongillo, and Christine M. Olson. 1998. “Food Insufficiency in the United States: Results From the Third National Health and Nutrition Examination Survey (NHANES III).” American Journal of Public Health 88:419–26. Andrews, Margaret, Mark Nord, Gary Bickel, and Steve Carlson. 2000. Household Food Security in the United States, 1999. Washington, DC: Economic Research Service, U.S. Department of Agriculture. Bartley, Mel. 1994. “Unemployment and Ill Health: Understanding the Relationship.” Journal of Epidemiology and Community Health 48:333–37. Bassuk, Ellen L., Linda F. Weinreb, John C. Buckner, Angela Browne, Amy Salomon, and Shari S. Bassuk. 1996. “The Characteristics and Needs of Sheltered Homeless and Low-Income Housed Mothers.” The Journal of the American Medical Association 276:640–46. Blazer, Dan. G., Ronald C. Kessler, Katherine A. McGonagle, and Marvin S. Swartz. 1994. “The Prevalence and Distribution of Major Depression in a National Community Sample: The National Comorbidity Survey.” American Journal of Psychiary 151:983–89. Bobo, Lawrence. 1995. Surveying Racial Discrimination: Analyses From a Multiethnic Labor Market. Working Paper No. 75, Russell Sage Foundation, Cambridge, MA. Booth, Cathryn L., Sandra K Mitchell, Kathryn E.



Barnard, and Susan J Spieker. 1989. “Development of Maternal Social Skills in Multiproblem Families: Effects on the MotherChild Relationship.” Developmental Psychology 25:403–12. Borjas, George J. 2001. Food Insecurity and Public Assistance. Working Paper No. 243, Joint Center for Poverty Research, Northwestern University/ University of Chicago, Chicago, IL. Bruce, Martha L., David T. Takeuchi, and Philip J. Leaf. 1991. “Poverty and Psychiatric Status.” Archives of General Psychiatry 48:470–74. Campbell, Cathy C. 1991. “Food Insecurity: A Nutritional Outcome or a Predictor Variable?” Journal of Nutrition 12:408–15. Christophar, S. P. and P. Peter Basiotis. 1992. “Dietary Intakes and Selected Characteristics of Women Ages 19–50 Years and their Children Ages 1–5 Years by Reported Perception of Food Sufficiency.” Journal of Nutrition Education 24:53–58. Corcoran, Mary E., Colleen M. Heflin, and Kristine Siefert. 1999. “Food Insufficiency and Material Hardship in Post-TANF Welfare Families.” Ohio State Law Journal 60:1395–1422. Dixon, Lori B., Marilyn A.Winkleby, and Kathy L. Radimer. 2001. “Dietary Intakes and Serum Nutrients Differ between Adults from FoodInsufficient and Food-Sufficient Families: Third National Health and Nutrition Examination Survey, 1988–1994.” Journal of Nutrition 131:1232–46. Gecas, Victor and Michael L. Schwalbe. 1983. “Beyond the Looking-Glass Self: Social Structure and Efficacy-Based Self-Esteem.” Social Psychology Quarterly 46:77–88. Heseker, H., W. Kübler, V. Pudel, and J. Westenhöffer. 1992. “Psychological Disorders as Early Symptoms of a Mild-to-Moderate Vitamin Deficiency.” Annals of the New York Academy of Sciences 669:352–57. Idler, Ellen L. and Yael Benyamini. 1997. “SelfRated Health and Mortality: A Review of Twenty-Seven Community Studies.” Journal of Health and Social Behavior 38:21–37 Idler, Ellen L. and Stanislav V. Kasl. 1995. “SelfRatings of Health: Do They Also Predict Change in Functional Ability?” Journal of Gerontology: Social Sciences 50B:S344–S353. Kendall, Anne, Christine M. Olson, and Edward A. Frongillo. 1996. “Relationship of Hunger and Food Insecurity to Food Availability and Consumption.” Journal of the American Dietetic Association 96:1019–24. Kessler, Ronald C., Gavin Andrews, Daniel K. Morczek, Bedirhan Ustun, and Hans-Ulrich Wittchen. 1998. “The World Health Organization Composite International Diagnostic Interview Short-Form (CIDI-SF).” International Journal of Methods in Psychiatric Research 7:171–85.

Kessler, Ronald C., Katherine A. McGonagle, Shanyang Zhao, Christopher B. Nelson, Michael Hughes, Suzann Eshleman, Hans-Ulrich Wittchen, and Kenneth S. Kendler. 1994. “Lifetime and 12-month Prevalence of DSM-IIIR Psychiatric Disorders in the United States: Results from the National Comorbidity Survey.” Archives of General Psychiatry 51:8–19. Kessler, Ronald C., Kristin D. Mickelson, and David R. Williams. 1999. “The Prevalence, Distribution, and Mental Health Correlates of Perceived Discrimination in the United States.” Journal of Health and Social Behavior 40:208–30. Krause, Neal and Thanh Van Tran. 1989. “Stress and Religious Involvement among Older Blacks.” Journals of Gerontology 44:S4–S13. Krieger, Nancy and Sally S. Zierler. 1996. “What Explains the Public’s Health? A Call for Epidemiologic Theory.” Epidemiology 7:107–09. Link, Bruce and Jo Phelan. 1995. “Social Conditions as Fundamental Causes of Disease.” Journal of Health and Social Behavior (Extra Issue):80–94. McEwen, Bruce S. 1998. “Protective and Damaging Effects of Stress Mediators.” New England Journal of Medicine 338:171–79. Mills, Bradford, Sundar Dorai-Raj, Everett Peterson, and Jeffrey Alwang. 2001. “Determinants of Food Stamp Program Exit.” Social Service Review 75:539–58. Mirowsky, John and Catherine E. Ross. 1986. “Social Patterns of Distress.” Annual Review of Sociology 12:23–45. Olson, Christine. 1999. “Nutrition and Health Outcomes Associated With Food Insecurity and Hunger.” Journal of Nutrition 129:521S–524S. Pearlin, Leonard I., Morton A. Lieberman, Elizabeth G. Menaghan, and Joseph T. Mullan. 1981. “The Stress Process.” Journal of Health and Social Behavior 22:337–56. Plewis, Ian. 1985. Analysing Change: Measurement and Explanation Using Longitudinal Data. New York: John Wiley and Sons. Quint, Janet C., Johannes M. Bos, and Denise F. Polit. 1997. New Chance: Final Report on a Comprehensive Program for Young Mothers in Poverty and their Children. New York: Manpower Demonstration Research Corporation. Quint, Janet, Rebecca Widon, with Lindsay Moore. 2001. “Post-TANF Food Stamp and Medicaid Benefits: Factors that Aid or Impede Their Receipt.” New York: Manpower Demonstration Research Corporation. Roberts, Gwenneth L., Gail M. Williams, Joan M. Lawrence, and Beverly Raphael. 1998. “How Does Domestic Violence Affect Women’s Mental Health?” Women and Health 28:117–29. Rodin, Judith. 1986. “Aging and Health: Effects of the Sense of Control.” Science 12:1271–76.

#1587—Jnl of Health and Social Behavior—Vol. 45 No. 2—45204-siefert FOOD INSUFFICIENCY AND PHYSICAL AND MENTAL HEALTH Rose, Donald. 1999. “Economic Determinants and Dietary Consequences of Food Insecurity in the United States.” The Journal of Nutrition 129 (2S Supplement):517-20S. Rose, Donald and Victor Oliveira. 1997. “Nutrient Intakes of Individuals from Food-Insufficient Households in the United States.” American Journal of Public Health 87:1956–61. Rothbaum, Fred, John R. Weisz, and Samuel S. Snyder. 1982. “Changing the World and Changing the Self: a Two-Process Model of Perceived Control.” Journal of Personality and Social Psychology 42:5–37. Scott, Richard I. and Cheryl A. Wehler. 1998. “Food Insecurity/Food Insufficiency: An Empirical Examination of Alternative Measures of Food Problems in Impoverished U.S. Households.” Discussion Paper No. 1176–98. Institute for Research on Poverty, Madison, WI. Sidel, Victor W. 1997. “Annotation: The Public Health Impact of Hunger.” American Journal of Public Health 87:1921–22. Siefert, Kristine, Phillip J. Bowman, Colleen M. Heflin, Sheldon Danziger, and David R. Williams. 2000. “Social and Environmental Predictors of Maternal Depression in Current and Recent Welfare Recipients.” American Journal of Orthopsychiatry 70:510–22. Siefert, Kristine, Colleen M. Heflin, Mary E. Corcoran, and David R. Williams. 2001. “Food Insufficiency and the Physical and Mental Health of Low-Income Women.” Women and Health 32:159–77. Somervell, Philip D., Philip J. Leaf, Myrna M.Weissman, Dan G. Blazer, and Martha L. Bruce. 1989. “The Prevalence of Major Depression in Black and White Adults in Five United States Communities.” American Journal of Epidemiology 130:725–35. Straus, Murray A. and Richard J. Gelles. 1986. “Societal Change and Change in Family Violence from 1975 to 1985 as Revealed by Two National Surveys.” Journal of Marriage and the Family 48:465–79. Tarasuk, Valerie S. and George H. Beaton. 1999. “Women’s Dietary Intakes in the Context of

185

Household Food Insecurity.” Journal of Nutrition 129:672–79. U.S. Bureau of the Census. 1998. Poverty in the United States: 1997. Current Population Reports, Series P-60, No. 201. Washington, DC: U.S. Government Printing Office. U.S. Council of Economic Advisors. 1999. Technical Report: The Effects of Welfare Policy and the Economic Expansion on Welfare Caseloads: An Update. Washington, DC: Executive Office of the President. Ware, John and Cathy D. Sherbourne. 1992. “The MOS 36-item Short-Form Health Survey (SF36): Conceptual Framework and Item Selection.” Medical Care 30:473–83. Wells, Kenneth B., Roland Sturm, Cathy D. Sherbourne, and Lisa S. Meredith. 1996. Caring for Depression: A RAND Study. Cambridge, MA: Harvard University Press. Williams, David R. 1997. “Race and Health: Basic Questions, Emerging Directions.” Annals of Epidemiology 7:322–33. Williams, David R., David T. Takeuchi, and Russell K. Adair. 1992. “Socioeconomic Status and Psychiatric Disorders among Blacks and Whites.” Social Forces 71:179–94. Williams, David R., Yan Yu, James S. Jackson, and Norman B. Anderson. 1997. “Racial Differences in Physical Mental Health: Socioeconomic Status, Stress and Discrimination.” Journal of Health Psychology 2:335–51. Wittchen, Hans-Ulrich. 1994. “Reliability and Validity Studies of the WHO-Composite International Diagnostic Review (CIDI): A Critical Review.” Journal of Psychiatric Research 28:57–84. World Health Organization. 1990. Composite International Diagnostic Review (CIDI, Version 1.0). Geneva, Switzerland: World Health Organization. Ziliak, James P., Craig Gunderson, and David Figlio. 2000. “Welfare Reform and Food Stamp Caseload Dynamics.” Discussion Paper No. 1215-00, Institute for Research on Poverty, University of Wisconsin, Madison, WI.

Kristine Siefert is Professor of Social Work and Associate Director of the NIMH Research Center on Poverty, Risk, and Mental Health at the University of Michigan. Her research focuses on social and environmental factors affecting the health of poor women and children and on racial/ethnic health disparities. Her current projects include studies of the impact of household food insufficiency on women’s health, the health of mothers under economic stress, and social-contextual determinants of oral health disparities. Colleen M. Heflin is Assistant Professor of Public Policy at the Martin School of Public Policy and Administration at the University of Kentucky. Her research focuses on welfare and public policy, social stratification, health inequality, and women and work. Current projects include an event history analysis of determinants of food stamp receipt in the welfare population, racial variation in levels of occupational sex segregation, causes and consequences of food insecurity, and barriers to employment among welfare recipients.



Mary E. Corcoran is Professor of Political Science, Public Policy, Social Work, and Women’s Studies, and Senior Associate Research Scientist in the Survey Research Center at the University of Michigan. Her research focuses on race and ethnicity-based differences in women’s wages and employment, on changes in welfare recipients’ employment as a result of welfare reform, and on the intergenerational transmission of economic inequality. David R. Williams is Harold W. Cruse Collegiate Professor of Sociology and Senior Research Scientist at the Survey Research Center, Institute for Social Research, at the University of Michigan. He is centrally interested in the determinants of socioeconomic and racial differences in physical and mental health. He is currently involved in projects examining discrimination and health, religious involvement and health, and the social distribution of psychiatric disorders in the United States and South Africa.