Disordered eating behaviors (DEB) are patterns of irregular or abnormal eating, often coupled with a component of psychological distress such as overevaluation of one’s weight or shape (1,2). They present in various ways, including chronic dieting, binge eating, compensatory behaviors (e.g., self-induced vomiting, over-exercising, use of laxatives or emetics, etc.), weight cycling, and intentional restriction of food intake (1); they may occur with other mental health comorbidities, including major depressive disorder or generalized anxiety disorder (3). DEB in adolescents carry more severe physical complications than adults (4); and they are associated with greater long-term body mass index (BMI) and waist circumference, and lower short- and long-term self-reported quality of life in youth and young adults youth and young adults (3,5–8).

DEB are twice as common in youth and young adults with type 1 diabetes (T1D) compared to those without diabetes (7.0% vs. 2.8%) (9) and more prevalent in youth and young adults with type 2 diabetes (T2D) than those with T1D (50.3% vs. 21.2%) (10). Specifically, DEB found in persons on insulin therapy are characterized by binge eating behavior followed by intentional withholding of insulin or other hypoglycemic drugs (with or without compensatory behaviors) in an attempt to induce negative energy balance (10,11). This may result in weight loss via hyperglycemia and glycosuria (12,13). Compared to people with diabetes and no DEB, those with DEB have ~1% higher hemoglobin A1c, twice the likelihood of diabetic ketoacidosis, and 1.5–2.5 fold increase in long-term complications of diabetes, including retinopathy, nephropathy and increased mortality (1, 4, 9, 12, 14–16).

Household food insecurity, an inability to access safe and nutritionally adequate food, is associated with higher depressive symptoms, stress, BMI, and probability of chronic diseases, and may further exacerbate risk of DEB (17–20). Household food insecurity is experienced as food scarcity, which likely predisposes individuals to compensatory food restrictive behaviors (19). Inconsistent food availability can trigger subsequent overconsumption (19). Altogether, this may result in cyclical eating, binge eating, food-related guilt, and weight self-stigma (17,21).

Furthermore, youth and young adults with household food insecurity and diabetes have additional stressors related to diabetes management such as regularly monitoring their blood glucose level and constantly adjusting their diet, physical activity, and diabetes medication to prevent or correct hyperglycemia and hypoglycemia, which could increase the risk of poor mental health (22), and thereby the risk for DEB. However, few studies have examined the relationship between household food insecurity and DEB among those with youth-onset T1D or T2D. Greater understanding of the relationship between household food insecurity and DEB among younger populations with diabetes could provide evidence to inform prevention and treatment of DEB. As such, the purpose of this study was to examine the association of household food insecurity levels and DEB among youth and young adults with youth-onset T1D or T2D. We hypothesized that higher levels of household food insecurity will be associated with greater levels of DEB in youth and young adults with T1D or T2D.

Participants were on average 21.5 ±5.1 years of age (26.2% aged 10–17 years, 73.8% aged 18–35 years), 54.2% female, 64.0% Non-Hispanic White (Table 1). Those with T1D were younger, less likely to be female, more likely to identify their race/ethnicity as Non-Hispanic White compared to participants with T2D. Participants had diabetes for an average of 11.2 ±3.3 years, 36.4% of them had a household income of at least $75,000, and 71.2% had private health insurance. Those with T1D had diabetes for a longer period, were more likely to live in a household with at least $75,000, and more likely to have private health insurance, compared with those with T2D.

The percentages of participants with marginal, low and very low food security were 11.6%, 9.9% and 7.7% in those with T1D and 20.0%, 18.0% and 12.4% in those with T2D (Table 1). The mean DEPS-R score was 13.5 ±9.5 for the total sample (23.2% with a score ≥20), 12.9 ±9.3 for those with T1D (20.8% with a score ≥20), and 17.5 ±9.6 for those with T2D (36.9% with a score ≥20). In households with very low food security, the mean DEPS-R scores for individuals with T1D and T2D were 18.6 ±11.8 and 21.1 ±11.7, respectively (Table 2). In households with high food security, the mean DEPS-R scores for individuals with T1D and T2D were 11.5 ±8.9 and 15.2 ±8.8, respectively.

Regardless of diabetes type, after adjusting for socio-demographics and duration of diabetes, youth and young adults with different levels of household food insecurity scored higher on the DEPS-R compared to those with high food security, the largest difference being with those with very low food security (5.9 points, 95% CI: 4.5, 7.3) (Table 3). Similar patterns were observed when stratifying by diabetes type, the largest differences in youth and young adults with T1D and T2D were with those experiencing very low food security who scored respectively 5.8 points (95% CI: 4.3, 7.4) and 6.6 points (95% CI: 3.3, 9.8) higher than their counterparts with high food security. Additional adjustments for BMI Z-score and/or depressive symptoms did not change the significance of the differences in DEPS-R scores, except for youth and young adults with T1D and marginal food security and youth and young adults with T2D and low food security where the confidence interval (in the model adjusting for both BMI Z-score and depressive symptoms) included the null value.

In a large sample of youth and young adults with youth-onset diabetes, we found that participants with different levels of household food insecurity had higher scores for DEB compared to those with high food security, in both those with T1D and T2D, even after adjustments for socio-demographics and diabetes duration. Larger effect sizes were observed among participants with more severe household food insecurity, indicating a dose-response relationship. These findings are similar to a study conducted among adolescents with T1D, which found that adolescent-reported food insecurity was associated with more DEB (34). Some differences with the present study concern the population (10–35 years of age with T1D or T2D vs. 10–17 years with T1D), the utilization of four food security levels rather than dichotomization which allows us to study how the severity of household food insecurity affects DEB, and the food security assessment (HFSSM reported by adults vs. Child Food Security Assessment questionnaire reported by adolescents). Parent report of children’s food insecurity experiences is the typical way to assess food security among youth (24).

Several potential mechanisms may explain the association between food insecurity and DEB in youth and young adults with diabetes. First, greater household food insecurity can influence multiple, overlapping psychosocial problems (35) such as emotional distress (36) and depressive symptoms (18) which are determinants of DEB (3,37). Youth and young adults with household food insecurity and diabetes have additional stressors on their psychosocial wellbeing concerning diabetes management which could increase vulnerability to development of poor mental health (22), and thus, higher risk for DEB. Second, due to the uncertainty in food availability inherent in food insecure households, meal structure and regularity can be hampered (38). Low meal structure, in turn, has been associated with subsequent development of disordered eating in children (39,40). This is consistent with our results by type of diabetes, where individuals with T2D tend to have a lower socioeconomic position (like US national data (41)) and are more likely to have household food insecurity and DEB compared to those with T1D.

There may be a synergistic effect of household food insecurity and DEB that can further inhibit glycemic regulation and long-term health outcomes in diabetes. Indeed, household food insecurity has been associated with elevated cortisol levels, insulin resistance and higher hemoglobin A1c levels in individuals living with diabetes (42,43). Moreover, DEB in persons with diabetes have also been associated with increased hemoglobin A1c and BMI, as well as greater adulthood weight and more frequent episodes of diabetic ketoacidosis (15,16,44). Elevated hemoglobin A1c in youth is associated with greater risk of microvascular complications in adulthood (45). Altogether, evidence is accumulating on the importance of addressing food insecurity and eating behaviors that may impair glucose regulation (e.g., low meal structure, insulin mismanagement, over-consumption of foods that increase glycosuria, food-related guilt that influences cyclical distress, etc.) and increase risk of long-term microvascular damage (46). Addressing these issues may require interventions at individual and structural levels because structural factors such as neighborhood food availability and access, food marketing policies, and government assistance play an important role in eating behaviors (47).

This study is innovative because it considers the level of household food insecurity and DEB in youth and young adults with youth-onset diabetes. This analysis benefited from the large size of the cohort (n = 2,669) and the use of validated survey tools despite the pitfall of self-reporting, and the fact that the DEPS-R has not been psychometrically tested for invariance by food security status. A limitation to this study is that characteristics which define household food insecurity and DEB may intersect, for example, cutting the size of meals or skipping meals. Nevertheless, the HFSSM places these experiences in the specific context that it is because there was not enough money for food. There were also differences in recall periods between the HFSSM (prior 12 months) and the DEPS-R (prior 1 month), which may have resulted in non-overlap of household food insecurity and DEB symptoms. Additionally, more than 30% of the observations have been imputed for CES-D and BMI Z-score. Nonetheless, the PMM produces precise and robust imputations (33), and results with or without adjustment for CES-D and BMI Z-score were similar. Temporality of associations found in this study cannot be established due to the cross-sectional nature of the data.

Household food insecurity levels were associated with DEB in youth and young adults with youth-onset T1D or T2D. Our findings suggest that addressing food insecurity may improve DEB in youth and young adults with diabetes. Our results are consistent with American Diabetes Association guidelines that emphasize the importance of screening for household food insecurity among individuals with diabetes (48). Thus, these results bolster evidence that clinicians and public health professionals should target individuals with diabetes and household food insecurity as being at higher risk for DEB. Regular screening for household food insecurity in clinical settings among youth and young adults with diabetes coupled with appropriate referrals to community and societal level assistance and social support tailored to the individual’s age and other characteristics can be beneficial.