Research Article
Determinants Of HealthFood Insecurity Is Associated With Higher Health Care Use And Costs Among Canadian Adults
- Fei Men ([email protected]) is a postdoctoral fellow in the Department of Nutritional Sciences at the University of Toronto, in Toronto, Ontario, Canada.
- Craig Gundersen is an ACES Distinguished Professor in the Department of Agricultural and Consumer Economics at the University of Illinois at Urbana-Champaign, in Urbana, Illinois.
- Marcelo L. Urquia is an associate professor in the Department of Community Health Sciences at the University of Manitoba, in Winnipeg, Manitoba, Canada. He is also a faculty member in Dalla Lana School of Public Health at the University of Toronto.
- Valerie Tarasuk is a professor in the Department of Nutritional Sciences at the University of Toronto.
Abstract
Food insecurity predicts poorer health, yet how it relates to health care use and costs in Canada remains understudied. Linking data from the Canadian Community Health Survey to hospital records and health care expenditure data, we examined the association of food insecurity with acute care hospitalization, same-day surgery, and acute care costs among Canadian adults, adjusting for sociodemographic characteristics. Compared with fully food-secure adults, marginally, moderately, and severely food-insecure adults presented 26 percent, 41 percent, and 69 percent higher odds of acute care admission and 15 percent, 15 percent, and 24 percent higher odds of having same-day surgery, respectively. Conditional on acute care admission, food-insecure adults stayed from 1.48 to 2.08 more days in the hospital and incurred $400–$565 more per person-year in acute care costs than their food-secure counterparts, with this excess cost representing 4.4 percent of total acute care costs. Programs reducing food insecurity, such as child benefits and public pensions, and policies enhancing access to outpatient care may lower health care use and costs.
Broadly defined, food insecurity is the “limited or uncertain availability of nutritionally adequate and safe foods or limited or uncertain ability to acquire acceptable foods in socially acceptable ways.”1 Population-level monitoring of food insecurity is more narrowly focused on the uncertainty of acquiring or inability to acquire sufficient food because of financial constraints. This experience-based measure of material deprivation is reported by one in eight Canadian households.2 In Canada, food insecurity is more prevalent among households with low income, low assets, and other characteristics associated with low resources and high costs for food access, such as single parenthood and residence in remote areas.3
Food insecurity has been linked to a variety of negative health outcomes such as higher risk for mental illness, physical chronic conditions, and premature deaths.4–9 Multiple mechanisms may explain the connection between food insecurity and poor health,9,10 including nutrient inadequacy,11 medication nonadherence,12 stress,13 social isolation,13 and constrained ability to manage chronic conditions.14 In Canada, where health care provision is universal, food-insecure adults can access health care services such as hospitalization and surgery free of charge, but they may experience financial barriers to disease prevention and management that require out-of-pocket expenditures, such as healthy diets and medications. Unlike health care services, prescription drugs in Canada are covered by a patchwork of public and private insurance, and many plans require copays and other out-of-pocket expenses at points of access. Inadequate self-care associated with food insecurity could complicate clinical conditions and hinder treatment and recovery. As a consequence, more advanced interventions such as hospitalizations and surgeries would be required for food-insecure adults, and once the patient is admitted to a hospital, prolonged stays and inflated costs would likely ensue. This may explain the previously observed association of food insecurity with greater health care use and expenditure in the Canadian province of Ontario and in the US.15–21
Food insecurity is key to understanding the role of social circumstances in health maintenance. However, the existing literature has primarily relied on self-reported rather than clinical measures of health.4–6 Moreover, in the Canadian context of universal health care, understanding of the implications of food insecurity for health services use and costs is still lacking: The few studies so far have focused on mental disorders,16 the province of Ontario,15–17 or the US.18–20 In addition, many past studies were limited by the use of a binary indicator for food insecurity,6,17–19 which may obscure important differences in health prediction that vary by severity of food insecurity, as illustrated elsewhere.4,5,15,16
Linking hospital records and cost estimates to data from the Canadian Community Health Survey (CCHS) 2005–17, we designed a population-based cohort study to examine the relationship between household food insecurity status and adults’ health care resource use, considering acute care hospital admissions and length-of-stay, same-day surgeries, main diagnoses associated with hospitalization and surgery, and health care expenditures associated with acute care hospitalization. We focused on hospitalizations and surgeries caused by illnesses and injuries and disregarded those caused by pregnancy and health consultation. The study was situated in the Canadian context, where the costs of acute care hospitalization and same-day surgery were fully covered by universal health insurance, thus minimizing the selection bias introduced by affordability.
Study Data And Methods
Study Design And Population
The CCHS is a cross-sectional survey conducted in two-year cycles, representing 98 percent of the Canadian noninstitutionalized population ages twelve and older. Since 2005 the survey has been used for population monitoring of household food insecurity, but the Household Food Security Survey Module was mandatory only in 2007–08, 2011–12, and 2017. Certain jurisdictions have chosen not to administer the module in cycles when it was optional (see online appendix exhibit 1).22
The Discharge Abstract Database is a national administrative database containing date and other clinical information on admissions and discharges between 1996 and 2017 from all acute care hospitals in Canada except Quebec.23,24 A “most responsible diagnosis” was assigned to every admission by trained professionals, using the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Canada (ICD-10-CA).25
The Canadian Institute for Health Information developed the Patient Cost Estimator in 2010 to estimate average acute care hospital costs (excluding physician fees) by jurisdiction and age group. Nunavut and the Northwest Territories were not included. The Patient Cost Estimator dollar amounts were calculated for the most common inpatient cases, aggregating acute care patients with similar clinical and resource use attributes.26
Because food insecurity status is measured for the prior twelve months, we linked CCHS data from the period 2005–17 for adults age eighteen or older to Discharge Abstract Database acute care admissions in the twelve months preceding the CCHS interview. Patient Cost Estimator data were not available before 2010; therefore, we limited our cost analyses to the interview years 2011–17, which correspond to the reference period for hospitalization (2010–17).
The initial data contained 667,285 CCHS respondents (appendix exhibit 2).22 We excluded Quebec for not having Discharge Abstract Database records and for having jurisdiction cycles that opted out of the Household Food Security Survey Module, respondents with missing data on food insecurity, and adolescents younger than age eighteen. The main analytic sample contained 403,620 adults. For the cost analysis, we used a subsample of 195,945 adults from the CCHS 2011–17 with matched cost information from the Patient Cost Estimator. Sampled adults tended to be economically better off and have healthier lifestyles than their excluded counterparts (appendix exhibit 3).22 Insofar as economic well-being and healthy lifestyle predict greater food security, our estimates would likely be downward biased.
Outcomes
The primary outcomes were all-cause acute care hospital admission, all-cause same-day surgery, and acute care costs. The secondary outcomes were readmission, year-cumulative length-of-stay (total number of days in acute care during the year preceding the CCHS interview), type-specific admission (urgent/emergent versus elective), cause-specific admission (eleven causes), cause-specific surgery (six causes), and high-cost users (top 5 percent of the expenditure distribution). We identified hospital admissions and same-day surgeries in the year before the CCHS (prior twelve months) to match with the reference period of food insecurity measurement.
The eleven causes of hospitalization and six causes of surgery were selected on the basis of common classification schemes by ICD-10 and case number for hospitalizations and surgeries contained in the sample (appendix exhibit 4).22,27 Statistics Canada requires a minimum of fifteen cases per food insecurity category per cause of hospitalization or surgery for the purpose of confidentiality. Causes of acute care admission and surgery resulting from pregnancy or health consultation were excluded from our analysis because of their potential link to proactive seeking of health care and the consequent selection bias. Patients with multiple hospitalization records during the tracing period could be matched to multiple causes. Acute care costs were adjusted for inflation.28 Consistent with prior research,17,19 the “high-cost user” indicator was created to identify the top 5 percent of the jurisdiction-specific cost distribution (including nonusers), taking into account the variation in health care use across Canadian jurisdictions.29,30
Exposure
The exposure of interest was twelve-month household food insecurity status. The variable was based on the validated eighteen-item questionnaire on a household’s access to food over the past twelve months, which was first created by the US Department of Agriculture and later adapted by Health Canada.2,31 Adults from households with no affirmative answer were deemed food secure, whereas those with any affirmative answer were labeled as marginally, moderately, and severely food insecure, depending on the number of affirmations on the survey (appendix exhibit 5).22
Covariates
On the basis of Ronald Andersen and John Newman’s theoretical framework on health services use,32 we hypothesized that individual demographic characteristics (sex and age) and lifestyle (smoking and alcohol consumption), household socioeconomic status (education, income, home ownership, and household composition), and contextual factors (jurisdictions and survey cycles) would jointly determine one’s health care use and cost. We also adjusted for individual baseline health by controlling for acute care hospitalization, or same-day surgery if the outcome was surgery, in the penultimate year before the CCHS interview (see technical appendix 1).22
Statistical Analysis
We first tabulated the sample characteristics by food insecurity status with person weights applied. We then used logistic regressions to estimate the adjusted odds ratios of food insecurity on acute care hospitalization and same-day surgery. We examined different causes of acute care hospitalization and same-day surgery, recognizing that socioeconomic disadvantage is more closely related to some conditions than to others.33,34 Limiting the data to patients who were ever hospitalized in the past year, we also estimated the odds of readmission and adjusted year-cumulative length-of-stay by food insecurity status.
We used two-part regression models to determine the relationship between food insecurity and health care expenditure, with logistic regression in the first step and linear regression with logged outcomes in the second step. Duan’s smearing estimate was used to convert logged estimates into actual dollar amounts.35 We estimated the odds of being high-cost users among people from the cost subsample. We further calculated the population-level annual excess acute care cost introduced by food insecurity by multiplying the average marginal costs estimated for each food insecurity category with their corresponding population weights.
We stratified the primary outcomes analyses by sex in light of the well-documented sex disparity in the way individual health reacts to social stressors.33,34 Analyses were performed with two-sided values, using Stata SE, version 15.1. The significance threshold was set to .
Limitations
This study had several limitations. First, food insecurity was assessed over a single twelve-month period only. Despite our effort to control for prior hospitalization and surgery, it remains probable that some unobserved causes could lead to food insecurity and hospitalization simultaneously. Indeed, multiple social determinants of health are predictors of food insecurity,3,36,37 some of which were unaccounted for here because of data constraints. Reverse causality is also possible: Receiving acute care could trigger financial struggles and mobility barriers, which in turn would increase food insecurity. Longitudinal data with repeated food security assessment would help approximate the causal effect of food insecurity on health care use and expenditure.
Second, our cost analysis was based on the Patient Cost Estimator’s population average estimates versus individual expenses. Although it served to approximate the difference in costs by food insecurity status, the actual dollar amounts may differ, with an uncertain direction of bias.
Third, we examined spending related to acute care only, because we lacked data for other relevant sources of costs such as same-day surgery and physician encounters. According to previous estimates from Ontario,15 hospitalization represented 28 percent of total health care costs, and food insecurity was associated with 34–144 percent higher total health care costs after accounting for all health services.
Last, the Discharge Abstract Database data do not include Quebec, the second-largest Canadian province and home to one-fifth of the Canadian population. Therefore, our sample is not nationally representative, and our results are not generalizable to Quebec. However, Quebec broadly resembles Canada in the cost of standard hospital stays, hospital readmission rate, and food insecurity rate;38,39 thus, adding Quebec to our sample would be unlikely to change our estimates substantively.
Study Results
Sample Characteristics
Of the population represented by the 403,620 adults in our sample, 3.74 percent, 5.25 percent, and 2.45 percent experienced marginal, moderate, and severe household food insecurity in the past twelve months, respectively. Compared with food-secure adults, those experiencing food insecurity were more likely to be women, younger, and less educated (appendix exhibit 6).22 Food-insecure adults were also less likely to be homeowners and more likely to be lone parents, current smokers, and admitted into acute care two years before the CCHS interview.
A total of 4.3 percent (n = 24,025) of sampled adults were admitted to an acute care hospital for at least one day in the prior twelve months, with an average length-of-stay of 8.2 days (data not shown). A total of 2.2 percent (n = 9,070) went through same-day surgery. The proportion hospitalized and the length-of-stay increased with the severity of food insecurity. Adults, on average, had acute care costs of $960 in the past year. Adults experiencing marginal or severe food insecurity had greater acute care costs than their food-secure counterparts. (For sample characteristics by food insecurity levels, see appendix exhibit 7.)22
Acute Care Hospitalization
Severity of food insecurity was positively associated with the likelihood of acute care admission in a dose-response fashion among adults after adjustment for confounders, with more severe food insecurity linked to a higher likelihood of hospital admission. Compared with food-secure adults, the adjusted odds of being hospitalized in the past twelve months was 26 percent (aOR: 1.26; 95% confidence interval: 1.17, 1.35), 41 percent (aOR: 1.41; 95% CI: 1.33, 1.50), and 69 percent (aOR: 1.69; 95% CI: 1.57, 1.83) higher for adults who were marginally, moderately, and severely food insecure, respectively, during this same period (exhibit 1; technical appendix 2).22 The associations with moderate and severe food insecurity were more pronounced among women than men (interaction for both).
Food secure (ref) | Marginally food insecure | Moderately food insecure | Severely food insecure | |
Overall | 1.00 | 1.26a**** | 1.41**** | 1.69**** |
By respondent’s sex | ||||
Male (n = 182,010) | 1.00 | 1.20*** | 1.31**** | 1.43**** |
Female (n = 221,610) | 1.00 | 1.28**** | 1.45**** | 1.83**** |
Among those hospitalized in the past year | ||||
Readmission (n = 24,025) | 1.00 | 0.98 | 1.24*** | 1.27*** |
Year-cumulative length-of-stay (days hospitalized; n = 24,025) | 0.00 | 1.50b*** | 1.48**** | 2.08**** |
Elective | 1.00 | 1.09 | 1.34**** | 1.38**** |
Urgent | 1.00 | 1.31**** | 1.45**** | 1.81**** |
Mental disorders | 1.00 | 1.65**** | 1.94**** | 2.94**** |
Symptoms with no diagnosis classified | 1.00 | 1.27** | 1.67**** | 2.15**** |
Injuries | 1.00 | 1.13 | 1.25*** | 1.38*** |
Digestive system diseases | 1.00 | 1.15 | 1.24*** | 1.33*** |
Circulatory system diseases | 1.00 | 1.33**** | 1.18** | 1.31** |
Respiratory system diseases | 1.00 | 1.24 | 1.48**** | 2.01**** |
Musculoskeletal diseases | 1.00 | 1.32*** | 1.45**** | 1.51**** |
Cancers | 1.00 | 1.06 | 1.20 | 1.05 |
Endocrine-metabolic system diseases | 1.00 | 1.05 | 1.50*** | 1.76**** |
Genitourinary system diseases | 1.00 | 1.20 | 1.64**** | 1.60**** |
Other miscellaneous causes | 1.00 | 1.27** | 1.43**** | 1.40*** |
Among people ever admitted in the past twelve months, moderate and severe food insecurity were associated with 24 percent (aOR: 1.24; 95% CI: 1.09, 1.41) and 27 percent (aOR: 1.27; 95% CI: 1.08, 1.49) higher odds of readmission in the same period (exhibit 1). Moreover, more severe food insecurity was associated with greater year-cumulative length-of-stay. Relative to the fully food-secure adults (predicted average length-of-stay: 8.41 days; 95% CI: 8.23, 8.60; data not shown), marginally, moderately, and severely food-insecure adults were associated with 1.50 (95 percent CI: 0.37, 2.62), 1.48 (95% CI: 0.62, 2.35), and 2.08 (95% CI: 0.89, 3.27) more days in acute care, respectively.
The association between food insecurity status and acute care admission was significant across types and causes of admissions in the overall sample. A dose-response relationship was found with both elective admissions and urgent/emergent ones, although marginal food insecurity was not associated with elective admissions. Moderate and severe food insecurity were associated with higher odds of acute care admission for all causes except cancers. Marginal food insecurity was associated with admissions for mental disorders, symptoms with no diagnosis classified, circulatory system diseases, and musculoskeletal diseases. Among the causes of hospitalization, mental disorders presented a particularly strong association with all levels of food insecurity.
Same-Day Surgery
Food insecurity was also associated with same-day surgery in a graded fashion (exhibit 2). Marginal, moderate, and severe food insecurity were associated with 15 percent (aOR: 1.15; 95% CI: 1.04, 1.26), 15 percent (aOR: 1.15; 95% CI: 1.06, 1.26), and 24 percent (aOR: 1.24; 95% CI: 1.11, 1.39) higher odds of having surgery, respectively (technical appendix 2).22 The association with marginal and moderate food insecurity was stronger for women than men (interaction for both).
Food secure (ref) | Marginally food insecure | Moderately food insecure | Severely food insecure | |
Overall | 1.00 | 1.15*** | 1.15**** | 1.24**** |
By respondent’s sex | ||||
Male (n = 182,010) | 1.00 | 1.05 | 1.07 | 1.23** |
Female (n = 221,610) | 1.00 | 1.19*** | 1.19*** | 1.22*** |
Symptoms with no diagnosis classified | 1.00 | 0.91 | 1.34** | 1.86**** |
Genitourinary system diseases | 1.00 | 1.45*** | 1.28** | 1.49*** |
Digestive system diseases | 1.00 | 1.03 | 1.14 | 1.24** |
Sense organs | 1.00 | 1.18 | 1.12 | 0.80 |
Cancers | 1.00 | 1.15 | 1.03 | 1.32* |
Other miscellaneous causes | 1.00 | 1.08 | 1.12 | 1.24** |
Severe food insecurity was positively associated with all causes of same-day surgery except surgery for sense organs (organs that respond to external stimuli by conveying impulses to the sensory nervous system; for example, cataract surgery on the eyes). Moderate food insecurity was associated with surgeries resulting from symptoms with no diagnosis classified and genitourinary system diseases. Marginal food insecurity was associated with surgery for genitourinary system diseases.
Acute Care Costs
Food-insecure adults had higher acute care costs than their food-secure counterparts, with costs increasing in a dose-response fashion as food insecurity becomes more severe (exhibit 3). The predicted average cost for food-secure adults was $1,285 (95% CI: 1,245, 1,323) over the past twelve months—the comparable figures for marginally, moderately, and severely food-insecure adults were $400 higher (95% CI: 164, 639), $535 higher (95% CI: 325, 747), and $565 higher (95% CI: 283, 849), respectively (technical appendix 3).22 No sex difference was found in predicted costs by food insecurity status. Total expenditure for acute care among adults from the sampled jurisdictions was estimated at $155 billion in 2011–17, of which 4.4 percent ($6.82 billion) could be considered as excess cost related to food insecurity. Compared with fully food-secure adults, those in marginal, moderate, and severe food insecurity had 28 percent (aOR: 1.28; 95% CI: 1.16, 1.42), 46 percent (aOR: 1.46; 95% CI: 1.34, 1.59), and 59 percent (aOR: 1.59; 95% CI: 1.42, 1.78) higher odds of being high-cost users, respectively (technical appendix 3).
Two-part regression on acute care expenditure | Food secure (ref) | Marginally food insecure | Moderately food insecure | Severely food insecure |
Adjusted odds ratio | 1.00 | 1.28**** | 1.46**** | 1.59**** |
Conditional acute care expenditure | $0.00 | $400**** | $535**** | $565**** |
Male (n = 87,790) | ||||
Adjusted odds ratio | 1.00 | 1.16 | 1.32**** | 1.26** |
Conditional acute care expenditure | $0.00 | $535** | $405** | $375 |
Female (n = 108,155) | ||||
Adjusted odds ratio | 1.00 | 1.36**** | 1.52**** | 1.77**** |
Conditional acute care expenditure | $0.00 | $300** | $545**** | $595**** |
Sensitivity Analyses
We performed two sensitivity analyses on all-cause acute care admission, length-of-stay, and all-cause surgery (appendix exhibit 8).22 Because food insecurity is rare among high-income households, we experimented with excluding adults with household income above $80,000 to eliminate potential biases from outliers. We also tried controlling for deaths in the postinterview year to rule out the impact of terminal illness on hospitalization. Those sensitivity tests yielded virtually the same results as our original ones. To explore the possibility of absent home support causing longer hospital stays, we further compared length-of-stay of unattached adults living alone with length-of-stay of couples with children but did not find a significant difference after confounders adjustment (linear coefficient: 0.125; ; data not shown).
Discussion
This study substantially advances understanding of the connection of household food insecurity with individual health, health care use, and public health costs in Canada. We found that severity of food insecurity was associated with not only greater likelihood of being hospitalized in acute care but also higher odds of readmission, longer stays, and higher expenditures once hospitalized. Dose-response relationships were evident, with more severe household food insecurity associated with higher health care use and costs. The associations with acute care hospitalization and same-day surgery were significant across nearly all primary causes.
Our findings are consistent with existing knowledge on the health correlates of food insecurity. Although prior studies have found food-insecure people to be at greater risk for falls,6 chronic diseases,4,5 and poor disease management,7,9 we show that these same people are also more likely to be hospitalized, particularly in an urgent/emergent manner, and to undergo same-day surgeries—all as possible consequences of inadequate disease management and prevention. That mental disorders were the cause of admission most strongly associated with food insecurity status is consistent with findings from an Ontario study on mental health resource use and the literature linking food insecurity to mental illnesses.4,5,15,16 Moreover, consistent with findings from the US,40 we found that the association between food insecurity and health care use was not confined to any single condition but, rather, was salient across nearly all major causes of hospitalization and surgery. The positive association of food insecurity with hospital admission, length-of-stay, and acute care cost was also in line with prior research with Ontario populations.15,16 Our findings suggest that food-insecure adults use more hospital resources and incur higher costs in acute care compared with their fully food-secure counterparts, with utilization rising systematically in relation to the severity of household food insecurity, possibly as a result of the greater complexity and severity of health problems experienced by adults in these conditions.
Policy Implications
Results from our study have important implications for public policy and practice. The association found with clinical versus self-reported measures of health is a substantial step toward establishing food insecurity as an independent social determinant of health. Preventing household food insecurity requires measures to address the resource constraints of vulnerable households. In Canada, policy interventions that improved income adequacy for low-income households, such as child benefits for families with children younger than age eighteen and public pensions for seniors ages sixty-five and older, have been associated with lower incidence and severity of household food insecurity.41–46 In the US, the Supplemental Nutrition Assistance Program has been shown to reduce food insecurity, although there is room for improvement in increasing benefits, reducing stigma, and simplifying the enrollment process.36,47–50 Our results suggest that adults’ health care use may also decrease with policy interventions that improve household food security, but more research is required to confirm this.
The strong intersection between food insecurity and health care use and expenditure reported here is evident under Canada’s universal health care system because acute care hospitalization and same-day surgery have no out-of-pocket charges, minimizing selection bias on cost. Moreover, unlike physician encounters and emergency department visits, hospitalization precludes selection bias associated with proactive health care seeking and reliably indicates the severity of conditions. Canada spent $253.5 billion on health care in 2018, of which 28.3 percent originated from hospitals.51 Our estimates of excess acute care resource use resulting from food insecurity suggest that measures to prevent food insecurity could reduce public health spending.
That adults experiencing more severe food insecurity use more acute care services is not a problem in itself. On the contrary, it may indicate proper treatment for those in need. However, policy makers need to reassess the availability and adequacy of primary care that may have prevented health deterioration in many cases, as well as the consequent use of acute care services.52 Difficulty affording prescription drugs is another major barrier to health maintenance and a possible reason for higher health care use and costs among food-insecure Canadians.12,53,54 A comprehensive national public medication insurance program (pharmacare) with universal exemption for out-of-pocket prescription expenses, which is currently under consideration by policy makers, could plausibly improve self-management of health, enhance equity in health care, and reduce health care use and costs.55,56
Similar to our findings, food insecurity is also associated with use of and total costs for health care in the US.18–21 However, unlike in the universal health care system in Canada, out-of-pocket expenses are common and are often substantial for health care services in the US,57 especially for people with no insurance coverage. Insofar as lower-income adults are less likely to have health insurance,58 out-of-pocket medical expenses could be high enough to deny many food-insecure adults access to proper care in the US, further jeopardizing their already-vulnerable health. Indeed, although higher out-of-pocket medical spending may aggravate food insecurity, food insecurity does not seem to affect this spending in the US,59 suggesting potential cost-related underuse of health care among food-insecure families.
Conclusion
This study marks an important step toward understanding the relationship of household food insecurity to individual health, health care use, and public health expenditure. Using clinical measures and a population-based sample from Canada, we show that severity of food insecurity was associated with greater health care use and spending in a dose-response fashion. Social policies that reduce food insecurity and enhance access to outpatient care may improve public health while lowering health care use and its associated costs.
ACKNOWLEDGMENTS
Preliminary results of this research were presented at the Canadian Research Data Centre Network National Conference in Halifax, Nova Scotia, Canada, October 24, 2019. This research was supported by Grant No. PJT 153260 from the Canadian Institutes of Health Research awarded to Valerie Tarasuk and Marcelo L. Urquia. The funding organization had no role in the design and conduct of the study; collection, management, analysis, and interpretation of data; or preparation, review, decision to submit for publication, or approval of the manuscript. This study has been approved by the Health Sciences Research Ethics Boards at the University of Toronto under Protocol No. 34032. Craig Gundersen received compensation from Feeding America and the Urban Institute. The authors thank the editor and the two anonymous reviewers for their constructive feedback. This research was conducted at RDC Toronto, a part of the Canadian Research Data Centre Network (CRDCN). This service is provided through the support of the University of Toronto, the Canadian Foundation for Innovation, the Canadian Institutes of Health Research, the Social Science and Humanity Research Council, and Statistics Canada. All views expressed in this work are the authors’ own.
NOTES
- 1 Department of Agriculture, Economic Research Service. Food security in the U.S.: measurement [Internet]. Washington (DC): USDA; 2019 Sep 4 [cited
2020 Jun 17 ]. Available from: https://www.ers.usda.gov/topics/food-nutrition-assistance/food-security-in-the-us/measurement.aspx#security Google Scholar - 2 . Household food insecurity in Canada, 2017–2018 [Internet]. Toronto: Research to Identify Policy Options to Reduce Food Insecurity (PROOF); 2020 [cited
2020 Jun 17 ]. Available from: https://proof.utoronto.ca/resources/proof-annual-reports/household-food-insecurity-in-canada-2017-2018/ Google Scholar - 3 . Geographic and socio-demographic predictors of household food insecurity in Canada, 2011–12. BMC Public Health. 2019;19(1):12. Crossref, Medline, Google Scholar
- 4 . The household food insecurity gradient and potential reductions in adverse population mental health outcomes in Canadian adults. SSM Popul Health. 2017;3:464–72. Crossref, Medline, Google Scholar
- 5 . Chronic physical and mental health conditions among adults may increase vulnerability to household food insecurity. J Nutr. 2013;143(11):1785–93. Crossref, Medline, Google Scholar
- 6 . Food insecurity associated with self-reported falls among Medicare Advantage members. Popul Health Manag. 2019;22(6):536–9. Crossref, Medline, Google Scholar
- 7 . Food insecurity is associated with hypoglycemia and poor diabetes self-management in a low-income sample with diabetes. J Health Care Poor Underserved. 2010;21(4):1227–33. Medline, Google Scholar
- 8 . Association between household food insecurity and mortality in Canada: a population-based retrospective cohort study. CMAJ. 2020;192(3):E53–60. Crossref, Medline, Google Scholar
- 9 . Hunger and socioeconomic disparities in chronic disease. N Engl J Med. 2010;363(1):6–9. Crossref, Medline, Google Scholar
- 10 . Food insecurity and health outcomes. Health Aff (Millwood). 2015;34(11):1830–9. Go to the article, Google Scholar
- 11 . Food insecurity is associated with nutrient inadequacies among Canadian adults and adolescents. J Nutr. 2008;138(3):604–12. Crossref, Medline, Google Scholar
- 12 . Prescription medication nonadherence associated with food insecurity: a population-based cross-sectional study. CMAJ Open. 2019;7(3):E590–7. Crossref, Medline, Google Scholar
- 13 . Food insecurity and mental illness: disproportionate impacts in the context of perceived stress and social isolation. Public Health. 2016;132:86–91. Crossref, Medline, Google Scholar
- 14 . The intersection between food insecurity and diabetes: a review. Curr Nutr Rep. 2014;3(4):324–32. Crossref, Medline, Google Scholar
- 15 . Association between household food insecurity and annual health care costs. CMAJ. 2015;187(14):E429–36. Crossref, Medline, Google Scholar
- 16 . The relation between food insecurity and mental health service utilization in Ontario. Can J Psychiatry. 2018;63(8):557–69. Crossref, Medline, Google Scholar
- 17 Looking beyond income and education: socioeconomic status gradients among future high-cost users of health care. Am J Prev Med. 2015;49(2):161–71. Crossref, Medline, Google Scholar
- 18 . Food insecurity and health care expenditures in the United States, 2011–2013. Health Serv Res. 2018;53(3):1600–20. Crossref, Medline, Google Scholar
- 19 . Food insecurity, healthcare utilization, and high cost: a longitudinal cohort study. Am J Manag Care. 2018;24(9):399–404. Medline, Google Scholar
- 20 . State-level and county-level estimates of health care costs associated with food insecurity. Prev Chronic Dis. 2019;16:E90. Crossref, Medline, Google Scholar
- 21 . Housing instability and food insecurity as barriers to health care among low-income Americans. J Gen Intern Med. 2006;21(1):71–7. Crossref, Medline, Google Scholar
- 22 To access the appendix, click on the Details tab of the article online.
- 23 Canadian Institute for Health Information. Discharge Abstract Database metadata (DAD) [Internet]. Ottawa (ON): CIHI; 2020 [cited
2020 Jul 10 ]. Available from: https://www.cihi.ca/en/discharge-abstract-database-metadata-dad Google Scholar - 24 Canadian Institute for Health Information. Data quality documentation, Discharge Abstract Database—multi-year information [Internet]. Ottawa: CIHI; 2012 [cited
2020 Jun 17 ]. Available from: https://www.cihi.ca/sites/default/files/document/dad_multi-year_en.pdf Google Scholar - 25 Canadian Institute for Health Information. Canadian coding standards for version 2012 ICD-10-CA and CCI [Internet]. Ottawa: CIHI; 2012 [cited
2020 Jun 17 ]. Available from: https://secure.cihi.ca/free_products/canadian_coding_standards_2012_e.pdf Google Scholar - 26 Canadian Institute for Health Information. Patient Cost Estimator [Internet]. Ottawa: CIHI; 2020 [cited
2020 Jun 17 ]. Available from: https://www.cihi.ca/en/patient-cost-estimator Google Scholar - 27 Government of Canada. Leading causes of death and hospitalization in Canada [Internet]. Ottawa: Public Health Agency of Canada; 2016 Mar 1 [cited
2020 Jun 17 ]. Available from: https://www.canada.ca/en/public-health/services/reports-publications/leading-causes-death-hospitalization-canada.html Google Scholar - 28 Statistics Canada. Table 18-10-0005-01 Consumer Price Index, annual average, not seasonally adjusted [Internet]. Ottawa: Statistics Canada; 2020 Apr [updated 2020 Jul 6; cited
2020 Jul 10 ]. Available from: https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1810000501 Google Scholar - 29 . Equity in health and health care in a decentralised context: evidence from Canada. Health Econ. 2008;17(3):377–92. Crossref, Medline, Google Scholar
- 30 . Does equity in healthcare use vary across Canadian provinces? Healthc Policy. 2008;3(4):83–99. Medline, Google Scholar
- 31 Health Canada, Office of Nutrition Policy and Promotion. Canadian Community Health Survey, Cycle 2.2, Nutrition (2004): income-related household food security in Canada [Internet]. Ottawa: Government of Canada; 2007 [cited
2020 Jun 17 ]. Available from: https://www.canada.ca/en/health-canada/services/food-nutrition/food-nutrition-surveillance/health-nutrition-surveys/canadian-community-health-survey-cchs/canadian-community-health-survey-cycle-2-2-nutrition-2004-income-related-household-food-security-canada-health-canada-2007.html Google Scholar - 32 . Societal and individual determinants of medical care utilization in the United States. Milbank Mem Fund Q Health Soc. 1973;51(1):95–124. Crossref, Medline, Google Scholar
- 33 . Social structure, stress, and mental health: competing conceptual and analytic models. Am Sociol Rev. 1991;56(2):166–78. Crossref, Google Scholar
- 34 . Becoming married and mental health: a longitudinal study of a cohort of young adults. J Marriage Fam. 1996;58(4):895–907. Crossref, Google Scholar
- 35 . Smearing estimate: a nonparametric retransformation method. J Am Stat Assoc. 1983;78(383):605–10. Crossref, Google Scholar
- 36 . Food insecurity research in the United States: where we have been and where we need to go. Appl Econ Perspect Policy. 2018;40(1):119–35. Crossref, Google Scholar
- 37 . Canada: a land of missed opportunity for addressing the social determinants of health. Health Policy. 2011;101(1):44–58. Crossref, Medline, Google Scholar
- 38 . Household food insecurity in Canada, 2012 [Internet]. Toronto: Research to Identify Policy Options to Reduce Food Insecurity (PROOF); 2014 Feb [last updated 2017 May 12; cited
2020 Jun 17 ]. Available from: https://proof.utoronto.ca/wp-content/uploads/2014/05/Household_Food_Insecurity_in_Canada-2012_ENG.pdf Google Scholar - 39 Canadian Institute for Health Information. An in-depth look at the Quebec health care system: explore Quebec health care using health indicators [Internet]. Ottawa (ON): CIHI; 2020 [cited
2020 Jun 17 ]. Available from: https://www.cihi.ca/en/an-in-depth-look-at-the-quebec-health-care-system Google Scholar - 40 . Food insecurity, chronic disease, and health among working-age adults [Internet]. Washington (DC): Department of Agriculture, Economic Research Service; 2017 Jul [cited
2020 Jun 17 ]. Available from: https://www.ers.usda.gov/webdocs/publications/84467/err-235_summary.pdf?v=0 Google Scholar - 41 . Money speaks: reductions in severe food insecurity follow the Canada Child Benefit. Prev Med. 2019;129:105876. Crossref, Medline, Google Scholar
- 42 . The impact of changes in social policies on household food insecurity in British Columbia, 2005–2012. Prev Med. 2016;93:151–8. Crossref, Medline, Google Scholar
- 43 . Reduction of food insecurity in low-income Canadian seniors as a likely impact of a guaranteed annual income. Can Public Policy. 2016;42(3):274–86. Crossref, Google Scholar
- 44 . An exploration of the unprecedented decline in the prevalence of household food insecurity in Newfoundland and Labrador, 2007–2012. Can Public Policy. 2015;41(3):191–206. Crossref, Google Scholar
- 45 . Household food insecurity in Ontario during a period of poverty reduction, 2005–2014. Can Public Policy. 2019;45(1):93–104. Crossref, Google Scholar
- 46 . A difference-in-differences approach to estimate the effect of income-supplementation on food insecurity. Prev Med. 2015;70:108–16. Crossref, Medline, Google Scholar
- 47 . Partial identification methods for evaluating food assistance programs: a case study of the causal impact of SNAP on food insecurity. Am J Agric Econ. 2017;99(4):875–93. Crossref, Google Scholar
- 48 . Food insecurity among households with children: the role of the state economic and policy context. Soc Serv Rev. 2017;91(4):691–732. Crossref, Google Scholar
- 49 . The right to food in the United States: the role of the Supplemental Nutrition Assistance Program (SNAP). Am J Agric Econ. 2019;101(5):1328–36. Crossref, Google Scholar
- 50 . An examination of the adequacy of Supplemental Nutrition Assistance Program (SNAP) benefit levels: impacts on food insecurity. Agric Resour Econ Rev. 2019;48(3):433–47. Crossref, Google Scholar
- 51 Canadian Institute for Health Information. National health expenditure trends, 1975 to 2018 [Internet]. Ottawa: CIHI; 2018 [cited
2020 Jun 17 ]. Available from: https://secure.cihi.ca/free_products/NHEX-trends-narrative-report-2018-en-web.pdf Google Scholar - 52 . Does continuity of care with a family physician reduce hospitalizations among older adults? J Health Serv Res Policy. 2006;11(4):196–201. Crossref, Medline, Google Scholar
- 53 . Medication non-adherence is associated with increased medical health care costs. Dig Dis Sci. 2008;53(4):1020–4. Crossref, Medline, Google Scholar
- 54 The consequences of patient charges for prescription drugs in Canada: a cross-sectional survey. CMAJ Open. 2018;6(1):E63–70. Crossref, Medline, Google Scholar
- 55 . Universal pharmacare in Canada: a prescription for equity in healthcare. Int J Health Policy Manag. 2020;9(3):91–5. Crossref, Medline, Google Scholar
- 56 . Prescription drug coverage in Canada: a review of the economic, policy, and political considerations for universal pharmacare. J Pharm Policy Pract. 2018;11:28. Crossref, Medline, Google Scholar
- 57 . Beyond the Numbers: household healthcare spending in 2014 [Internet]. Washington (DC): Department of Commerce, Bureau of Labor Statistics; 2016 Aug [cited
2020 Jun 17 ]. Available from: https://www.bls.gov/opub/btn/volume-5/pdf/household-healthcare-spending-in-2014.pdf Google Scholar - 58 . Income, poverty, and health insurance coverage in the United States: 2005 [Internet]. Washington (DC): Census Bureau; 2006 Aug [cited
2020 Jun 17 ]. (Current Population Reports No. P60-231). Available from: https://www.census.gov/prod/2006pubs/p60-231.pdf Google Scholar - 59 . Food insecurity and out-of-pocket medical expenditures: competing basic needs? Fam Consum Sci Res J. 2010;39(2):137–51. Crossref, Google Scholar