This study used data from the Korea Welfare Panel Study (KOWEPS) survey conducted by the Korea Institute for Health and Social Affairs and Seoul National University. The KOWEPS has been conducted annually since 2006 as face-to-face interviews. It targets households living across the country and covers a wide range of topics, such as health and health care, economic activity status, income, and consumption. In particular, these surveys have significantly covered households with incomes below 60% of the standard median income, and 100% of household heads have been covered by either national health insurance or government benefits [21]. Thus, data from the KOWEPS are appropriate for confirming depression and economic pressure on household heads. In this study, we used 2020 data as the most recent year of published data. A total of 6,029 households were surveyed in the 2020 data, including heads of households aged between 22 and 98 years. This study targeted households with heads younger than 65 years (2,853 households). The aim of this study was to identify depression among working-age household heads who face significant economic burdens in supporting their families. After filtering and removing missing data, the final sample comprised 2,839 households, corresponding to 2,839 working-age household heads. Detailed information about these surveys and the KOWEPS data is publicly available on the survey organization’s website [22]. Because the KOWEPS data used in this study are publicly available, ethics approval from the institutional review boards of the authors’ affiliated institutions was not required.

The KOWEPS considers participant depression based on an 11-item version of the Center for Epidemiologic Studies Depression Scale and includes lack of appetite, unhappy state, depression, difficulty with activities of daily living, sleeplessness, loneliness, complaints, unfriendly treatment, sadness, dislike, and demoralization. This scale is a shortened version of the 20-item scale developed by Radloff [23]. This shortened version has been used extensively in recent research and has been found to be preferable to the original 20-item version because it reduced the likelihood of false positives, particularly in older adults [24]. In addition, results from previous studies demonstrated that the 11-item scale was highly correlated (Pearson correlation coefficient, r=0.95) with the full version [25]. There were four responses (0=“almost never”; 1=“some of the time”; 2=“much of the time”; and 3=“most of the time”) for each item, and participants indicated the corresponding score reflecting their experience of the previous week. The depression index was calculated as the sum of the scores on the 11 items. Participants with a higher depression index experienced more severe depression. In this study, to avoid data dispersion, we used the threshold proposed by the KOWEPS survey organization to define depression status. Accordingly, participants with a depression index less than 16 were considered non-depressed; otherwise, they were considered depressed. This approach is consistent with those of previous studies [21,25].

Demographic and health parameters of the household heads were considered in the research criteria. Demographics included sex (male and female), age group (22–54 years and 55–64 years or older, corresponding to the prime working-age group and mature working-age group, respectively, based on the classification of the South Korean Age structure [26]), educational status (middle school or lower, high school, and college or higher), marital status (married; unmarried including single mothers; and others including bereaved, divorced, and separated), and type of employment (wage earners, employers, and unemployed). The health parameters were self-rated health (good or not good), disability status (none, moderate, or severe), and the presence of chronic disease (yes or no).

Household parameters played an important role in our study to indicate the depression of household heads. Household-related variables were defined based on KOWEPS survey information. These included household income classification (general and low, based on 60% of the standard median income before receiving government benefits), number of household members, presence of disabled members (yes or no), presence of the elderly aged 65 years or older (yes or no), and presence of children under 15 years of age (yes or no). Age determination for the elderly (65 years or older) and children (under 15 years) was based on the classification of the South Korea Age structure [26].

We determined the prevalence of depression among household heads based on demographic characteristics, health status of the household head, and household parameters. Bivariate analyses were employed to determine participant characteristics and estimate the prevalence of depression. Fisher’s exact test was used to calculate P-values for the prevalence of depression across participant characteristics and household parameters [27]. A P-value less than 0.05 represented a significant difference.

A multiple logistic regression model was used to evaluate the influence of these factors on the presence of depression [28]. In this model, the depression status of the household head was the dependent variable, while the predictors were demographics, health parameters of the household head, and household parameters. The Wald test was used to calculate the P-value for parameter estimation, and the likelihood-ratio chi-square (LRC) test was used to assess model fit [29]. The “not depressed” category of depression status was designated as the reference for the model. The odds ratios (ORs) were obtained by exponentiating the regression coefficients and the corresponding 95% confidence intervals (95% CIs). The OR represents the relative risk ratio of depression status associated with a one-level change in the respective predictor. All analyses were conducted using the R software ver. 4.2.1 (The R Foundation for Statistical Computing, Vienna, Austria; https://www.r-project.org/).

We included 2,839 household heads with a mean age of 49.23 years (standard deviation=10.24 years; range, 22–64 years), of whom 2,254 (approximately 79.4%) were male. The overall prevalence of depression was 11.69%; however, 19.83% of the female household heads had depression, which was more than double the prevalence among males (9.58%). Participants of prime working-age (22–54 years) comprised 63.75% of the study population, but the prevalence of depression among them (10.17%) was significantly lower than in the mature working-age population (14.38%). Considering educational status, household heads with a college education or higher accounted for the largest proportion (47.52%), while those with a middle school education or lower accounted for 14.93%. The prevalence of depression decreased sharply with an increase in education level: 19.58% among household heads with middle school education or lower, 13.04% among those with high school education, and 8.15% among those with college education or higher. The prevalence of depression varied sharply by marital status, from 8.23% among married couples to 15.61% among unmarried people (including single mothers) and up to 20.73% among those bereaved, separated, or divorced. With respect to the employment type, the lowest prevalence of depression was among wage earners (8.53%) and the highest was among unemployed household heads (29.73%). Those who rated their health as “not good” had a high prevalence of depression (38.22%). Similarly, the depression rate among those with severe disabilities was 35.64%. The characteristics of the household heads and prevalence of depression are shown in Table 1.

Depression rates among household heads according to household parameters are shown in Table 2. Low-income households accounted for 85.45% of all households. The depression rate among the heads of low-income households was approximately 31%, which was more than 3.6 times higher than the rate among the general-income households. The depression rate among the household heads gradually decreased as the number of family members increased. Specifically, for single-member households, the depression rate among household heads was 17.2%, which decreased to 14.51% for two-member households, 10.7% for three-member households, and 7.79% for households with more than three members. This was similar in households with and without children. Specifically, the depression rate among heads of households with children was low (7.27%), whereas it was 13.51% for households without children. In households with disabled members, the depression rate among household heads (19.6%) was nearly twice as that among households without disabled members.

Figure 1 illustrates the sex-differentiated impacts of demographics, health status, and household parameters on the prevalence of depression among household heads. In general, the rate of depression was significantly higher among female household heads than male household heads. However, in the group of household heads with a severe disability, the prevalence of depression was significantly lower among females than among males (25.81% versus over 40%, respectively) and even lower than the prevalence in females with a moderate disability (39.13%). There was no significant difference in depression between males and females in households with children.

The results of the multiple logistic regression analysis are presented in Table 3. The small P-value from the LRC test (<0.00001) demonstrates the adequacy and robustness of the model.

The results obtained from multiple logistic regression analysis showed the varying influences of factors on depression among household heads. All factors had a significant effect on depression (P<0.05). The relative risk of depression was 1.21 times higher among female household heads than among the male household heads. Unmarried household heads (including single mothers) had a relative risk of depression that was 1.32 times higher than married household heads and 1.83 times higher than household heads who were bereaved, divorced, or separated. Similarly, the relative risk of depression was 1.71 times higher for unemployed household heads than for wage earners. The relative risk of depression was 3.23 times higher in those with poor health status than in those with good health, and 2.45 times more in those with severe disabilities than in those without disabilities. Those with chronic diseases had a higher relative risk of depression (1.63 times) than those without. Household income had the strongest influence on depression among household heads. The relative risk of depression among household heads with low household incomes was 2.18 times higher than the risk among household heads with general household incomes. The relative risk of depression among household heads decreased with the number of family members or family members who were children but increased in households with members who were disabled or elderly.