Subjects and general characteristics
We used data from the 2014 KNHANES for 4349 subjects aged 19 years and older. Individuals who did not participate in the health behavior interviews and 24-h dietary recall tests and those who reported eating fewer than 500 kcal or more than 5000 kcal of daily total caloric intake were excluded to minimize biases (Fig. 1). All the KNHANES questionnaires used in this study were approved by the Institutional Review Board of the Korea Centers for Disease Control and Prevention (approval number: 2013-12EXP-03-5C).
The general characteristics of the subjects were analyzed according to sex, age, marital status, residential area, occupational status, educational level, household income, and average family size. Age was classified as follows: 19 to 29, 30 to 49, 50 to 64, 65 to 74, and 75 years and older. Residential areas were classified as “city” and “rural.” The following classifications were also used: household income (low, middle-low, middle-high, and high), educational level (middle school or less, high school or less, and college degree or more), and occupational status (employed and unemployed).
Subjects with depression were selected based on their answers to the Patient Health Questionnaire (PHQ)-9 part of the health survey in the 2014 KNHANES. The PHQ-9 is a self-rated diagnostic tool for depression listed in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition [21,22,23,24]. It is a widely used and well-validated measure for monitoring depressive symptoms . Han et al.  validated a Korean version of the PHQ-9 that is used in clinical trials and medical research settings to assess depression. Participants were asked, “How often have you been bothered by any of the following symptoms over the previous two weeks?” The PHQ-9 uses nine items to measure the severity of depressive symptoms: little pleasure in activities, feelings of hopelessness or feeling down, sleep disturbances (trouble falling asleep or staying asleep or sleeping too much), feeling tired or having little energy, changes in appetite (poor appetite or overeating), feelings of guilt or worthlessness, trouble concentrating, feeling lethargic or fidgety, and feeling suicidal . Each of the nine items is rated on a four-point scale of 0 (not at all), 1 (several days), 2 (more than half the days), and 3 (every day), and the answers are summed to provide the total PHQ-9 score. Based on previous studies, subjects with PHQ-9 score ≥ 10 (of 27 points) were defined as having depression in this study [23, 24, 26, 27].
Health behavior measures
The health behaviors assessed in the current study were smoking status (non-smoker, ex-smoker, or current smoker), drinking status (≥ 4 times/week, 2–3 times/week, 1–4 times/month, or < 1 time/month), stress status (very much, somewhat, a little, and rarely), and exercise status (< 1 day/week, 1–2 days/week, 3–4 days/week, and ≥ 5 days/week). Based on body mass index (BMI kilograms/square meter), the weight status was divided into four groups: underweight (< 18.5), normal (18.5–23.0), overweight (23.0–25.0), and obese (≥ 25.0).
Dietary behavior measures
The number and type of meals consumed (breakfast, lunch, and dinner) and the average frequency per week of eating out (≥ 2 times/day = 14, once per day = 7, 5–6 times/week = 5.5, 3–4 times/week = 3.5, 1–2 times/week = 1.5, 1–3 times/month = 0.32, rarely = 0) were analyzed. Four groups of food security status were identified according to the dietary survey included in the KNHANES data : “food secure” (able to meet essential food and non-food needs for all family members without depletion of assets), “mildly insecure” (minimally adequate food consumption but unable to afford some essential non-food expenditures without depletion of assets), “moderately insecure ” (marginally able to meet minimum food needs because of insufficient money), and “severely insecure” (often not enough food to eat because of insufficient money, has large food consumption gaps).
Food intake was assessed using the 24-h recall method. Based on previous studies [29, 30], individual food intake was categorized into 18 food groups: total food, cereals and grain products, potatoes and starches, legumes and their products, seeds and nuts, sugars and sweets, vegetables, fruits, seaweeds, fish and shellfish, meat, poultry and its products, oils and fats, seasonings, mushrooms, eggs, milk and dairy products, beverages, and other foods. Vegetables were classified into two categories of non-salted or non-starchy vegetables (excluding pickled and salted vegetables, starchy vegetables, and juice) and salted vegetables (including pickles, kimchi, and fermented vegetables). Fruit was categorized as fresh fruit (excluding jams, sweetened fruits, and juices) and sweetened fruit (including jams).
The KNHANES was conducted using a nationally representative estimate of the Korean population based on a multistage, stratified, cluster sampling method. The statistical analyses in this study were performed by adopting stratification, clustering, and sample weight variables using SAS version 9.4 statistical software (SAS Institute, Cary, NC, USA). Cronbach’s α was used to determine the internal consistency of the PHQ-9 items. For general characteristics and health and dietary behaviors according to depression status, the results are reported as frequency and weighted percent from the frequency analysis. Chi-squared testing was used to identify significant differences among categorical variables. For the PHQ-9 items of age, family size, average eating-out frequency per week, average meal frequency per day, BMI, average sleep time, and food and nutrient intake according to depression status, means and standard errors were calculated using the surveymean procedure. Significant differences were verified using an unadjusted t test. A generalized linear model was used after adjusting for sex, age, and energy intake with the surveyreg procedure. The correlation between vegetable/fruit intake and depression was determined using a logistic regression analysis with quartile groups (Q1, Q2, Q3, Q4) of non-salted vegetables and fresh fruit intake as the independent variable and depression status (depression, 1; normal, 0) as the dependent variable using the surveylogistic procedure. The amount of non-salted vegetables and fresh fruits eaten was divided into quartiles, and the lowest quartile (smallest consumption of vegetables and fruits) was used as the reference category. The results are presented as odds ratios (ORs) with 95% confidence intervals (CIs). The logistic regression analysis was performed after adjusting for sex, age, energy intake, smoking, drinking, exercise, stress, snacks, breakfast, marital status, eating-out frequency, food security, and household income in stages.