Association Between Weight Control Failure and Suicidal Ideation in Overweight and Obese Adults

A Cross-sectional Study

Yeong Jun Ju; Kyu-Tae Han; Tae-Hoon Lee; Woorim Kim; Jeong Hun Park; Eun-Cheol Park


BMC Public Health. 2016;16(259) 

In This Article


Study Population

For this cross-sectional study, we used raw data from the Korea National Health and Nutrition Examination Surveys (KNHANES) conducted in 2008–2012 (4th and 5th KNHANES); these are cross-sectional surveys that have been conducted annually by the Korea Centers for Disease Control and Prevention (KCDC). These are cross-sectional surveys with study populations from multistage, stratified area probability samples of civilian non-institutionalized Korean households by geographic area, age, and gender groups. This survey is composed of three parts—a health interview, health examination, and nutrition survey—all of which were performed by trained medical staff and dieticians. We used data from all three parts. These data were collected from a total of 45,811 participants during 2008–2012 (8058 in 2008, 8518 in 2009, 8958 in 2010, 10,533 in 2011, 9744 in 2012). The overall response rates were 74.3 % in 2008, 79.2 % in 2009, 81.9 % in 2010, 80.4 % in 2011, 80.8 % in 2012. In our analysis, the subjects were divided into two subgroups according to BMI: 23–24.9 kg/m2 (overweight) or ≥25 kg/m2 (obese). Any respondents who did not provide data on BMI, suicidal ideation, education level, income, household composition, marital status, moderate physical activity, stress awareness, alcohol consumption, or who were under the age of 40 years were excluded from the study (see details in Fig. 1). In addition, we excluded those who did not undergo weight control attempts or who attempted to gain weight to investigate the success or failure of weight control efforts in the participants. A total of 6621 eligible participants were included in the present study (2439 overweight; 4182 obese). Meanwhile, considering that sex affects mental health, the present study analyzed men and women separately (overweight: 1116 males, 1323 females; obese: 2081 males, 2101 females).[4,21] The KNHANES data are openly available at the KNHANES website:

Figure 1.

Flow diagram of the study participants

Ethical approval for this study was granted by the institutional review board (IRB) of the KCDC Seoul, Korea (IRB #: 2008-04EXP-01-C; 2009-01CON-03-2C; 2010-02CON-21-C; 2011-02CON-06-C; 2012-01EXP-01-2C).


Suicidal ideation. We used data from a self-reported questionnaire, in which suicidal ideation was assessed by a single item: "Have you ever seriously thought about suicide in the past year?" Responses were rated on a 2-point scale: yes or no. Despite limitations due to the question's simplicity, previous studies have successfully used this method based on KNHANES data.[4,22,23]

Weight Control Failure. To assess the success or failure of weight control efforts, we used data from a self-reported questionnaire regarding changes in body weight and weight control efforts. Change in body weight was assessed by the single question, "Have you experienced a change in body weight in comparison with the last year?" Responses were rated on a 3-point scale: weight maintenance, weight gain, or weight loss. Weight gain is composed of three parts: increases of 3–6 kg, 6–10 kg, or 10 kg or more. Weight loss was composed of three parts: decreases of 3–6 kg, 6–10 kg, or 10 kg or more. Weight control efforts were assessed by the single question, "Have you ever tried to control your weight voluntarily in the last year?" Responses were rated on a 4-point scale: made efforts for weight loss, made efforts for weight maintenance, made efforts for weight gain, did not make efforts for weight control. Finally, we classified weight control failure as "yes" or "no". Those who experienced weight gain during weight loss efforts were categorized as "yes". Those who experienced weight loss during weight loss efforts or weight maintenance during weight maintenance efforts were classified as "no".

Criteria of Obesity. The World Health Organization Regional Office for the Western Pacific Region recommends defining obesity in Asians as those with a BMI ≥25 kg/m2. Korea officially uses this definition when calculating the prevalence of obesity in Korea. Overweight in Asians is defined as a BMI of 23–24.9 kg/m2.[24] Therefore, our study used this definition.

Covariates. We included age, sex, income, educational level, economic activity, marital status, household composition, smoking status, alcohol consumption, stress awareness, depressive symptoms, perceived body image, perceived health status, moderate physical activity, menopause (only for females) and survey year as covariates. Age was categorized into four groups, starting with 40 years of age. Income status was classified as "low", "middle", or "high". Economic activity was classified as "employed" or "unemployed". Household composition was assessed by the question "What is your household composition?" This variable was categorized as "one-generation household" (single-person household), "two-generation household" (couple + children, single parent + children, couple + parents, couple + single parent, grandparents + grandchildren, single grandparent + grandchildren), and "three-generation household" (couple + parents + children, couple + single parent + children). Stress awareness was classified as "high" or "low". Depressive mood was assessed by duration of the depressive mood during the past 2 weeks (present, absent). Subjective body perception was assessed by the question "How would you categorize your current body shape?" (very thin, slightly thin, normal, slightly fat, or very fat). We categorized this as "slim and normal" (very thin, slightly thin, or normal) or "fat" (slightly fat or very fat). Moderate physical activity was classified as whether respondents performed moderate physical activity for 30 min per session more than five times per week. Menopause status was assessed as "yes" (experienced menopause) or "no" (not applicable or yet to experience menopause).

Statistical Analysis

We first examined the distribution of each categorical variable. The Chi-squared test was used to calculate the frequencies and percentages of the variables and to identify significant differences between groups. The analyses for incidence of suicidal ideation by sex were also performed. In addition, to produce an unbiased national estimate, a sample weight was assigned for the participating individuals to represent the Korean population. The sampling weight were calculated by accounting for the complex survey design, survey nonresponse, and post-stratification.[25] Next, multivariable logistic regression analysis was used to examine the association between weight control failure and suicidal ideation while controlling for potential confounding variables such as age, sex, income, educational level, economic activity, marital status, household composition, alcohol consumption, stress awareness, depression mood, subjective body image, perceive health status, moderate physical activity, menopause (only females) and survey year. In addition, the fit of the model was assessed using the Hosmer-Lemeshow goodness-of-fit test; well-fitting logistic models have a nonsignificant goodness-of-fit.[26] Multicollinearity was tested using a variance inflation factor (VIF), which provided an index to measure how much the variance of an estimated regression coefficient increased due to collinearity. If the value of the variance inflation factor exceeded 10, a model was regarded as indicating collinearity. Finally, subgroup analyses were performed to evaluate the association between weight control failure and suicidal ideation according to income, household composition, or menopause. All of the analyses were performed using SAS 9.4 software (SAS Institute, Cary, NC, USA) and the statistical significance level was set at p-value < 0.05.