Seligman HK, Jacobs EA, Lopez A, Sarkar U, Tschann J, Fernandez A. Food Insecurity and Hypoglycemia Among Safety Net Patients With Diabetes. Arch Intern Med. 2011;171(13):1204-1206. doi:10.1001/archinternmed.2011.287
Author Affiliations: Division of General Internal Medicine (Drs Seligman, Sarkar, and Fernandez and Ms Lopez) and Center for Vulnerable Populations (Drs Seligman and Sarkar and Ms Lopez), University of California, San Francisco, San Francisco General Hospital; Center for Obesity Assessment, Study, and Treatment (Drs Seligman and Tschann) and Department of Psychiatry (Dr Tschann), University of California, San Francisco; and Department of Medicine and Health Innovation Program, University of Wisconsin-Madison (Dr Jacobs).
In 2008, 1 in every 7 US households was food insecure or at risk of going hungry because of an inability to afford food.1 Food insecurity generally occurs cyclically and episodically, with periods of food adequacy followed by food scarcity; the average food-insecure household experiences 7 episodes of food scarcity annually.2 We hypothesized that the cyclic nature of food insecurity and its associated fluctuations in dietary intake would predispose patients with diabetes to wide variations in blood glucose levels, including clinically significant hypoglycemia.
We conducted a cross-sectional survey and medical chart review of patients with type 2 diabetes mellitus receiving care in community health clinics in San Francisco, California, and Chicago, Illinois. All participants were 18 years or older, fluent in English or Spanish language, and had a documented diagnosis of diabetes. We measured food insecurity using the 6-item US Household Food Security Survey Module.3 We measured our primary outcome of severe hypoglycemia using the same question used in landmark diabetes studies: “In the past year, how many times have you had a severe low blood sugar reaction, such as passing out or needing help to treat the reaction?”4 Categorical response options included 0, 1 to 3, 4 to 6, 7 to 11, or 12 or more times in the previous year. We dichotomized severe hypoglycemia at 4 or more times based on response distribution. To validate self-reported hypoglycemia, we compared self-reported to clinical diagnoses of hypoglycemia (International Classification of Diseases, Ninth Revision code 251*) for a subset of 237 participants from 1 health system. We found a high correlation (P = .002) and good between-measure agreement (κ = 0.7).
We determined whether food insecurity was an independent risk factor for 4 or more episodes of severe hypoglycemia using a multivariate logistic regression model including sociodemographic characteristics differing by P < .20 in univariate analysis (age, race/ethnicity, smoking, English language proficiency, income, and education) and known or hypothesized risk factors for hypoglycemia (insulin use, renal disease, adherence to medication and blood glucose testing, body mass index, comorbid disease burden, hypoglycemia knowledge, problem alcohol use, and hemoglobin A1c [HbA1c] level). Participants who had experienced a severe hypoglycemic episode were asked whether their blood glucose “had ever gotten too low because you couldn't afford to buy enough food.”
The study was approved by the institutional review boards of the sponsoring institutions, and participants provided written informed consent.
Of the 782 eligible patients approached, 711 (91%) completed interviews. Participants were 53% Latino, 30% African American, and 17% white. Almost half of participants (n = 325 [46%]) were food insecure. Food-insecure participants were significantly younger (mean, 53 vs 56 years; P < .001), had lower household incomes (81% vs 70% reporting <$25 000 annually; P < .001), and were more likely to smoke (34% vs 17%; P < .001) than food-secure participants. Mean HbA1c level was higher among food insecure than food-secure participants (8.54% vs 8.09%; P = .007).
Almost 28% of participants (n = 197) reported 1 or more severe hypoglycemic episode in the previous year, and 9.4% of participants reported 4 or more episodes. Participants who were food insecure were significantly more likely than those who were food secure to report 4 or more episodes (12.6% vs 6.7%; odds ratio, 2.00 [95% confidence interval, 1.19-3.35]). This association remained significant after adjusting for demographic factors and other risk factors for hypoglycemia with an odds ratio of 2.95 (95% confidence interval, 1.48-5.91). Food insecurity, problem alcohol use, comorbid illness burden, and body mass index were significantly associated with hypoglycemia in the fully adjusted model (Table). Sensitivity analyses dichotomizing the hypoglycemia variable at 7 or more episodes or 12 or more episodes did not alter the results.
Among participants reporting hypoglycemic episodes, those who were food insecure were significantly more likely to attribute hypoglycemia to the inability to afford food (43.2% vs 6.8%; P < .001).
In our 2-city sample, we found that almost half of patients with diabetes seeking care in urban, safety-net community health clinics are at risk of hunger because of the inability to afford food. This food insecurity is a significant risk factor for frequent episodes of severe hypoglycemia, even after adjusting for other risk factors for hypoglycemia. We observed a higher prevalence of severe hypoglycemia (28%) than that reported in another study of community-dwelling adults (11%).5 Food insecurity may partially explain this elevated risk.
Hypoglycemia is one of the most common adverse drug events leading to emergency department visits and hospitalizations,6 even though most hypoglycemic episodes are never brought to medical attention. Depending on severity and duration, hypoglycemia may reduce quality of life and cause traumatic accidents, cognitive dysfunction, and death. Ensuring that low-income patients using hypoglycemic medications have reliable access to food is therefore an important patient safety issue that clinicians in safety-net settings should consider when patients report frequent hypoglycemia. Clinicians should consider “food insecure” any patient who agrees that in the last year “we worried whether our food would run out before we had money to buy more” or “the food we bought just didn't last and we didn't have money to get more.”7
Although other studies have also suggested that food insecurity is associated with poorer overall glycemic control as measured by high HbA1c levels, to our knowledge, ours is the first to report on hypoglycemic episodes. The higher HbA1c levels we observed in food-insecure compared with food-secure participants suggests that tighter glycemic control does not explain the increased hypoglycemia risk associated with food insecurity. Although it may appear contradictory that food insecurity is associated with both hypoglycemia and hyperglycemia, there are several potential explanations. First, lengthy episodes of food adequacy, during which blood glucose levels may run high, are often followed by brief episodes of food scarcity.8 Second, clinicians may relax glycemic targets when patients encounter frequent or severe hypoglycemic episodes. Such relaxation of glycemic targets may be appropriate for some patients. However, screening low-income patients who report frequent hypoglycemic episodes for food insecurity and providing assistance in obtaining reliable access to food may help reduce socioeconomic disparities in diabetes care and outcomes.
Correspondence: Dr Seligman, Division of General Internal Medicine, Department of Medicine, University of California, San Francisco, PO Box 1364, San Francisco, CA 94143-1364 (email@example.com).
Author Contributions:Study concept and design: Seligman, Jacobs, and Fernandez. Acquisition of data: Seligman, Jacobs, Lopez, and Fernandez. Analysis and interpretation of data: Seligman, Jacobs, Sarkar, Tschann, and Fernandez. Drafting of the manuscript: Seligman and Fernandez. Critical revision of the manuscript for important intellectual content: Seligman, Jacobs, Lopez, Sarkar, Tschann, and Fernandez. Statistical analysis: Seligman and Tschann. Obtained funding: Seligman, Jacobs, and Fernandez. Administrative, technical, and material support: Jacobs, Lopez, and Fernandez. Study supervision: Jacobs and Fernandez.
Financial Disclosure: None reported.
Funding/Support: This project was supported by funding from the Russell Sage Foundation and The Commonwealth Fund. Dr Seligman is supported by NIH/NCRR/OD UCSF-CTSI grant KL2 RR024130 from the National Institutes of Health. Dr Fernandez receives additional support from an Arnold P. Gold Foundation Professorship.
Role of the Sponsors: The funders did not participate in the design and conduct of the study; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript.
Disclaimer: The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Previous Presentations: This work has been presented in abstract form at the following meetings: Clinical and Translational Research and Education Meeting, ACRT/SCTS Joint Annual Meeting; April 6, 2010; Washington, DC; Society of General Internal Medicine Meeting, May 1, 2010, Minneapolis, Minnesota; and Centers for Disease Control and Prevention's Diabetes Translation Conference; April 13-15, 2010; Kansas City, Missouri.