Page 19 - IJPS-9-1

International Journal of
Population Studies Food insecurity severity and depression

2.2. Major depression 9–11 grade, high school graduate/General Educational
th
Our outcome of interest was major depression. We Development, some college, and college graduate or above.
constructed this variable using the Patient Health Access to care is a binary yes or no response based on the
Questionnaire-9 (PHQ-9), a nine-question depression question, “Have you seen a mental health professional in
module with each question ranging from a frequency of the past year?”.
zero (not at all) to three (nearly every day) in the past 2.5. Statistical analysis
12 months. The aggregate score had a range of 0–27 and we
followed guidance on classifying individuals with major A Pearson Chi-square test was conducted to test for an
depression using a cutoff of 10 or greater, which has a high association between adult food security and depression.
sensitivity and specificity of 88% (Kroenke et al., 2001). A weighted multiple logistic regression was run to
evaluate the relationship between food insecurity and
2.3. Food insecurity severity depression after adjusting for covariates. The NHANES
Our main predictor variable was the severity of adult sampling weights were calculated based on the probability
food insecurity derived from a 10-question NHANES of selection at each stage of the sampling process, which
algorithm that queried adult households with no included stratification, clustering, and multistage
children. The NHANES uses the standardized 10-item sampling. To make sure that the sample was representative
food security module developed by the USDA, which has of the population of interest, the sampling weights were
been extensively tested for validity and reliability (Bickel adjusted for nonresponse by demographic factors such as
et al., 2000). The 10 questions in the module assess three age, sex, race/ethnicity, and poverty status. This ensures
situations: (1) Anxiety or perception that the food budget or that the data are representative of the U.S. civilian, non-
food supply was inadequate, (2) perceptions that the food institutionalized population. We used IBM SPSS statistical
eaten was inadequate in quality, and (3) reported instances software version 28 and SAS Studio 3.81 for all statistical
of reduced food intake or consequences of reduced intake. analyses with a 0.05 significance level and two-tailed tests.
Based on USDA methodology, the NHANES algorithm
classifies adults into four categories: (1) Full food security 3. Results
(had no problems, or anxiety about, consistently accessing 3.1. Sample characteristics
adequate food), (2) marginal food security (had problems
at times, or anxiety about, accessing adequate food, but In our study sample of 5,856 U.S. adults, the average age of
participants was 49.9 years (SD = 18.8) and the majority
the quality, variety, and quantity of their food intake were of participants identified as female (51.5%) (Table 1).
not substantially reduced), (3) low food security (reduced
the quality, variety, and desirability of their diets, but the The average ratio of family income to poverty was 2.52
quantity of food intake and normal eating patterns were (SD = 1.61), indicating mid-level socioeconomic status.
not substantially disrupted), and (4) very low food security Most of the participants identified as non-Hispanic White
(at times during the year, eating patterns were disrupted, (34.7%), while those who identified as other Hispanic
and food intake reduced due to lack of money and other made up the lowest proportion of study participants.
resources for food) (Bickel et al., 2000). The highest proportion of study participants reported
an education level of some college (32.0%). Participants
2.4. Covariates who reported only having a 9–11 grade or <9 grade
th
th
The regression model was adjusted for the following education level represented the lowest proportion of our
covariates: Age, income, race/Hispanic origin, sex, sample. When examining measures of health, 89.7% of
education level, and access to care. Age was a continuous participants reported that they had not seen a mental
variable ranging from 18 and top-coded at 80 years to health professional in the last year. In addition, 63.3% of
protect respondent anonymity. For income, we specifically participants were categorized as experiencing full food
analyzed the ratio of family income to poverty, which was security, while only 9.6% of participants experienced very
calculated by dividing family or individual income by the low food security. About 8% of the NHANES participants
poverty guidelines specific to the survey year. This variable met the criteria for major depression based on the PHQ-9.
is continuous with a range of zero and top-coded to five 3.2. Bivariate results
to protect respondent anonymity. Race/Hispanic origin is
nominal, and responses included Mexican American, other There was a statistically significant relationship
Hispanic, non-Hispanic White, non-Hispanic Black, and between food insecurity severity and major depression
2
other races. Sex includes male and female. Education level (ꭓ [3] = 197.9, p < 0.001) (Table 2). In general, more
is ordinal and includes the following responses: <9 grade, severe food insecurity trended with a higher prevalence
th
Volume 9 Issue 1 (2023) 13 https://doi.org/10.36922/ijps.435