Discussion

The limited access to healthy foods and poor disease management are important in the risk of diseases. Adults with food insecurity have been known to be more likely to have barriers to disease management and a higher risk of chronic diseases than adults with food security. Their passive disease management would exacerbate the risk of diseases, but studies on the combined effects of disease management and food insecurity on health outcomes are scarce. The present study found that among disease management-related factors, taking medications to treat diseases was not different by food insecurity status, but adults with food insecurity had higher unmet health-care needs and mental health counselling than adults with food security. The difference in disease management between adults with food security and food insecurity had a different impact on mental health and physical health outcomes.

Similar results have been shown in other studies in which food insecurity was associated with a lack of disease management and the two coexisting factors increased adverse effects on health^{(Reference Afulani, Herman and Coleman-Jensen11–Reference Kushel, Gupta and Gee14,Reference Seligman, Davis and Schillinger31,Reference Chan, DeMelo and Gingras32)} . People on low incomes with food insecurity have high rates of poor access to care and are more likely to delay care when needed, tending to result in increased risk of obesity and diabetes^{(Reference Kushel, Gupta and Gee14,Reference Chan, DeMelo and Gingras32)} . In the adult participants in the National Health Interview Survey, food-insecure people were less likely to use cost-related medication and suffered more from chronic conditions^{(Reference Berkowitz, Seligman and Choudhry12,Reference Herman, Afulani and Coleman-Jensen13)} . Among adults with diabetes, food insecurity was associated with poor diabetes self-management and increased risk of hypoglycaemia^{(Reference Seligman, Davis and Schillinger31)}. On the other hand, a recent Canadian study reported that food insecurity was associated with increased health-care use and higher costs; the contrast with previous studies has been explained by different access to health insurance in the context of Canada’s universal health-care system^{(Reference Tarasuk, Cheng and de Oliveira33)}.

The possible aetiological mechanism underlying the association between food insecurity and physical or mental health outcomes has been proposed in previous studies^{(Reference Laraia2,Reference Seligman, Laraia and Kushel3,Reference Seligman, Davis and Schillinger31,Reference Seligman and Schillinger34)} . Households with food insecurity could be more likely to experience episodic food shortages, thereby following an abnormal cyclic pattern of food consumption and metabolism alteration, and would be more dependent on inexpensive and high-energy-density foods. These negative experiences would also promote a stress response resulting in a preference for and intake of highly palatable foods that are energy dense and which can promote diet-induced obesity or insulin resistance^{(Reference Laraia2,Reference Seligman, Laraia and Kushel3,Reference Seligman, Davis and Schillinger31,Reference Seligman and Schillinger34)} . In addition, the stress resulting from food shortages, constrained dietary options and complementary coping behaviours among food-insecure persons with low socio-economic status would result in psychological distress and poor mental health^{(Reference Chung, Kim and Kwak6–Reference Davison and Kaplan8)}. The pathways between food insecurity and health outcomes might be influenced by disease management^{(Reference Afulani, Herman and Coleman-Jensen11–Reference Kushel, Gupta and Gee14,Reference Seligman, Davis and Schillinger31,Reference Chan, DeMelo and Gingras32)} . When food insecurity results in the development of chronic diseases, health-care expenditure and the time required for managing disease will increase. These circumstances put further strain on economic resources, exacerbating food insecurity. In this way, a cyclic relationship between food insecurity, poor health outcomes and disease management arises^{(Reference Seligman and Schillinger34)}. Thus, understanding the role of disease management in household food insecurity and the development of disease could be important for interventions targeting modifiable factors such as disease management.

In the present study, higher unmet health-care needs was an important risk factor for health problems in adults with food insecurity. Several studies have reported the major reasons for unmet health-care needs as financial burdens, limited availability of services or functional limitations in adults with food insecurity^{(Reference Berkowitz, Seligman and Choudhry12–Reference Berkowitz, Meigs and DeWalt15,Reference Kim, Lee and Moon35,Reference Srinivasan and Pooler36)} . Additionally, a recent Korean study using Community Healthy Survey data found that use of health-care services to address mental health problems was positively associated with a high educational level^{(Reference Jang, Lee and Kim37)}. The study explained that the reasons could be that higher educated adults could have more opportunities to obtain information on access to mental health services and less stigma of what others would think about receiving mental health services. The findings implied that an increased accessibility to mental health services and increased public awareness about the services are important for vulnerable populations, including adults with food insecurity.

Despite the importance of mental health counselling, ironically, adults with food insecurity who underwent mental health counselling were more likely to have mental health problems than those who did not undergo counselling. This unexpected result could be explained by reverse causal inference because of the cross-sectional nature of the data. Given that persistently needy individuals seek more help, an at-risk population with mental health problems might be more likely to undergo mental health counselling. The findings also reflect that the current services could not be adequate to manage these problems. Several studies reported that legal frameworks have been established to implement community-based mental health services, but professional staff and budget are lacking^{(Reference Jung, Che and Chun38,Reference You and Lee39)} . In addition, negative attitudes towards the services, distrust of the service effects, or lack of money and time could be the reasons for the non-persistent use of mental health services among the vulnerable population^{(Reference Lee and Song40,Reference Lee41)} . Therefore, for the low-income mental health high-risk group, the mental health service system might be improved first in terms of resource planning and management, as well as promotion to raise the awareness of services.

For physical health outcomes, the prevalence of obesity and underweight was different between adults with food security and insecurity, with a higher prevalence of underweight and obesity in men and women with food insecurity, respectively. The physical health outcomes were the worst in adults with food insecurity who did not meet health-care needs or did not receive nutrition education or counselling in the present study. In recent times, obesity is a major health issue in the disadvantaged population, particularly among children and women in Korea^{(Reference Chun, Ryu and Park10,Reference Yoon and Kwon42)} . The disparity would be partially mediated by a lower intake of healthy foods, such as vegetables and fruits^{(Reference Hong, Kim and Kim18–Reference Kim, Shin and Shim21)}. The present study showed that insufficient disease management, such as unmet health-care needs and low participation in nutrition education or counselling, could also affect the disparity of physical health problems. The health-care system should be more affordable and accessible to adults with food insecurity to reduce the inequality of the health-care burden which leads to health disparity.

Generally, in the present study, the gender difference in the relationship between food insecurity, disease management and health outcomes was not pronounced, suggesting that the potential influence of food insecurity and disease management on health is similar for men and women. Unexpectedly, both men and women were more likely to be obese or underweight regardless of food insecurity when they were trying to control weight compared with those who were not. This might be explained by reverse causality, like the result for the mental health counselling.

The present study has some limitations that should be addressed in future studies. First, the study was based on a cross-sectional survey; hence, only an association was observed and causality could not be confirmed clearly. Second, the mental health outcomes were assessed using a single-item self-reported questionnaire. In terms of the question for suicidal ideation, the single-item self-reported assessment had the possibility of increased risk of misclassification^{(Reference Millner, Lee and Nock43)}. On the other hand, several Western studies have found that a single-item self-reported question for depressive experience may be useful to identify depressive patients^{(Reference Reme and Eriksen44–Reference Maboney, Drinka and Abler46)}. However, one study among Korean older adults reported that a single question could substantially underestimate depression^{(Reference Kim, Chun and Kim47)}. Therefore, the prevalence of mental health outcomes in the present study may be underestimated and further studies using questionnaires fully valid for mental health outcomes are necessary. Third, the status of household food insecurity was applied to individuals to measure food insecurity. There is a possibility that an individual’s experience of food insecurity and that of family members are likely to differ depending on the allocation of resources within the household^{(Reference Bickel, Nord and Price1)}. However, the study population is adult members who are likely to be similarly food insecure^{(Reference Bickel, Nord and Price1)}, so that individual food insecurity measures applied from the household level measures might be suitable. Finally, we could not include possible confounding factors that were not available in the data set, such as substance use which could be an effect modifier of mental health in relation with food insecurity^{(Reference Davison, Holloway and Gondara48)}. Despite several limitations, the strength of the present study is the use of a nationally representative sample showing that intervention in health-care management may require a policy to further reduce the inequality in health problems due to food insecurity.