Few studies have estimated the economic burden of chronic diseases (e.g., type 2 diabetes, cardiovascular diseases, cancers) attributable to unhealthy eating. In this study, we estimated the economic burden of chronic disease attributable to not meeting Canadian food recommendations. We first obtained chronic disease risk estimates for intakes of both protective (1. vegetables; 2. fruit; 3. whole grains; 4. milk; 5. nuts and seeds) and harmful (6. processed meat; 7. red meat; 8. sugar-sweetened beverages) foods from the Global Burden of Disease Study, and food intakes from the 2004 Canadian Community Health Survey 24-hour dietary recalls (n = 33,932 respondents). We then calculated population attributable fractions (PAFs) for all relevant food-chronic disease combinations by age and sex groups. These PAFs were then mathematically combined for each disease for each age and sex group. We then estimated attributable costs by multiplying these combined PAFs with estimated 2014 annual direct health care (hospital, drug, physician) and indirect (human capital approach) costs for each disease. We found that not meeting recommendations for the eight foods was responsible for CAD$13.8 billion/year (direct health care: CAD$5.1 billion, indirect: CAD$8.7 billion). Nuts and seeds and whole grains were the top cost contributors rather than vegetables and fruit. Our findings suggest that unhealthy eating constitutes a tremendous economic burden to Canada that is similar in magnitude to the burden of smoking and larger than that of physical inactivity which were estimated using similar approaches. A status quo in promotion of healthy eating will allow this burden to continue. Interventions to reduce the health and economic burden of unhealthy eating in Canada may be more effective if they are broad in focus and include promotion of nuts and seeds and whole grains along with vegetables and fruit rather than have a narrow focus such as primarily on vegetables and fruit.
Citation: Lieffers JRL, Ekwaru JP, Ohinmaa A, Veugelers PJ (2018) The economic burden of not meeting food recommendations in Canada: The cost of doing nothing. PLoS ONE 13(4): e0196333. https://doi.org/10.1371/journal.pone.0196333
Editor: David Meyre, McMaster University, CANADA
Received: December 22, 2017; Accepted: April 11, 2018; Published: April 27, 2018
Copyright: © 2018 Lieffers et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data used for this study are third-party data available from Statistics Canada through the Research Data Centres (RDC) program. The Statistics Act prohibits Statistics Canada from releasing any information that could possibly identify an individual. Only researchers working on research projects approved by Statistics Canada and who have been sworn in under the Statistics Act as 'deemed employees' will receive permission to access the data. All data analysis procedures must be completed in the RDC facility where confidential data are stored on an internal computer network that is protected from outside access. In addition, all outputs that researchers wish to physically remove from the RDC facilities must undergo a careful and strict vetting procedure by a Statistics Canada's employee. For more information about the RDC program, please visit the following website: https://www.statcan.gc.ca/eng/rdc/index. Researchers who do wish to access the data used for this study can do so if they submit an application to Statistics Canada. For more information on Research Data Centre policies, please contact STATCAN.MADHOOU-DAMSSOBC.STATCAN@canada.ca. Questions or comments on the privacy practices at Statistics Canada may be directed to: Departmental Privacy Officer, Statistics Canada, R.H. Coats Building, 2nd floor, 100 Tunney's Pasture Driveway, Ottawa, Ontario K1A 0T6, Telephone: 613-951-0466.
Funding: This research was funded through the Collaborative Research and Innovation Opportunities (CRIO) Team program from Alberta Innovates (AI) (grant number 201300671 to PJV). PJV holds a Canada Research Chair in Population Health, an Alberta Research Chair in Nutrition and Disease Prevention, and an Alberta Innovates Health Scholarship. All interpretations and opinions in the present study are those of the authors. This research was also supported by funds to the Canadian Research Data Centre Network (CRDCN) from the Social Science and Humanities Research Council (SSHRC), the Canadian Institute for Health Research (CIHR), the Canadian Foundation for Innovation (CFI) and Statistics Canada. Although the research and analysis are based on data from Statistics Canada, the opinions expressed do not represent the views of Statistics Canada or the Canadian Research Data Centre Network (CRDCN).
Competing interests: The authors have declared that no competing interests exist.
In Canada, chronic diseases (e.g., type 2 diabetes, cardiovascular diseases, cancers) were the leading causes of death in 2013  and are responsible for billions of dollars in direct health care (i.e., hospitalization, drug, and physician costs) and indirect (i.e., lost productivity due to disability, morbidity, and premature mortality) costs each year based on the Economic Burden of Illness (EBIC) study . With a population that is both aging and living longer, decreasing the burden of chronic diseases should be a priority to alleviate future tension on the health care system and improve overall productivity.
Several factors are known to cause chronic diseases including some that are non-modifiable (e.g., age, genetics). However, up to 80% of type 2 diabetes and cardiovascular diseases, and over 1/3 of cancers are caused by four modifiable behavioral risk factors: tobacco use, harmful alcohol use, physical inactivity, and unhealthy eating . With various competing acute care demands and limited resources available for primary prevention, public health decision makers have difficult choices to make when allocating funds to achieve the best benefits in terms of avoiding health care costs and improving productivity.
Public health decision makers often consult cost of illness studies to understand the economic burden associated with different chronic disease risk factors . To date, in Canada, estimates of the economic burden associated with different chronic disease risk factors have mostly focused on smoking [5–13], excess body weight [5–7,9,13–19], physical inactivity [5–7,9,13,14,20,21], and alcohol consumption [7,12,13,22]. For example, in 2013, the direct health care and indirect costs attributable to excess body weight, smoking, alcohol consumption, and physical inactivity were estimated to amount to CAD$23.5 billion, CAD$19.5 billion, CAD$10.6 billion, and CAD$9.3 billion, respectively . Fewer studies have been conducted to estimate the economic burden associated with unhealthy eating despite many Canadians not meeting established food recommendations . One possible reason there are few estimates is because this calculation is more complex due to both multiple risk factors and exposure levels.
One approach commonly used to estimate the economic burden associated with unhealthy eating is to quantify the impact of a single dietary factor [13,24–32]. For example, we recently estimated that the inadequate intake of vegetables and fruit costs Canadian society CAD$3.3 billion/year in direct health care and indirect costs . Abdullah et al  found that if all adults in Canada met fibre recommendations, there would be potential savings in direct health care and indirect costs of CAD$1.3 billion/year for cardiovascular diseases and CAD$718.8 million/year for type 2 diabetes. Jones et al  reported that intake of sugar-sweetened beverages in Canada will cost CAD$33.7 billion over the next 25 years. However, because these studies focused on single dietary factors they underestimate the overall economic burden of chronic disease associated with unhealthy eating.
There have only been a few studies that estimated the overall economic burden of unhealthy eating. These studies have primarily been macroeconomic and ecologic in nature which do not consider food recommendations and specific dietary intakes. For example, in the United Kingdom, Scarborough et al  estimated that poor diets were responsible for £5.8 billion in National Health Services expenditures in 2006–07. Using a similar approach, Cancer Care Ontario and Public Health Ontario estimated that the direct health care costs associated with unhealthy eating in the Canadian province of Ontario were CAD$2.9 billion in 2011 . Frazão et al  estimated that the economic burden of chronic diseases associated with unhealthy eating in the United States in 1995 was USD$70.9 billion in direct health care and indirect costs. Using an adapted approach , Health Canada estimated that the costs of unhealthy eating to the Canadian society in 1998 were $6.6 billion ($1.3 billion due to direct health care costs) .
The aim of this study was to estimate the economic burden of chronic disease attributable to unhealthy eating using both food recommendations and intakes from Canada. Unhealthy eating in this study was defined as not meeting established food recommendations.
Materials and methods
This economic burden of not meeting food recommendations pertains to the mismatch between the amount of certain foods that is recommended and the amount that is actually consumed in Canada. It represents an estimate of costs that could potentially be avoided if all Canadians would fully comply with all food recommendations. We estimated the economic burden of not meeting food recommendations using a prevalence-based approach. This approach involved estimating the economic burden of chronic disease in a given period of time (i.e., one year) regardless of disease onset. Our estimations used a bottom up approach; this involved estimating the economic burden of different chronic diseases attributable to different foods for specific age and sex groups and then combining these estimates.
To obtain economic burden estimates, we first calculated population attributable fractions (PAF). Population attributable fractions use both the relative risk of disease due to a certain exposure and the distribution of the risk factor in a specific population (e.g., specific age and sex group) to estimate the fraction of disease cases that would not occur should the exposure be eliminated from the entire population (i.e., if everyone followed food recommendations). We then multiplied the PAF values by the overall economic costs associated with relevant chronic diseases to determine the estimated economic burden attributable to unhealthy eating.
We used multiple steps in our analyses which included: a) selection of foods for inclusion and extraction of dose-response relative risks; b) analysis of food consumption data; c) calculation of population attributable fractions; and d) estimation of attributable direct health care and indirect costs. These four steps are described in detail below.
Ethical approval for this study was provided by the University of Alberta Research Ethics Board (Pro00073196). The Canadian Community Health Survey data were accessed through Statistics Canada’s Research Data Centers (RDC) program.
We used foods in our analysis that were identified by the 2015 Global Burden of Disease Study (GBD) as having convincing or probable evidence against ≥1 chronic diseases . The GBD summarized epidemiological evidence on various food-chronic disease combinations to identify those that either have convincing or probable evidence according to World Cancer Research Fund criteria for grading evidence . We included the following eight foods in our analyses that had convincing or probable evidence against ≥1 chronic diseases: fruit not including juice, non-starchy vegetables, whole grains, nuts and seeds, fluid milk, red meat, processed meat, and sugar-sweetened beverages (SSB). We extracted dose-response relative risks for each of these foods for chronic disease morbidity from the 2015 GBD  with the exception of SSB which were extracted from the 2013 GBD (note: the relative risks for SSBs were estimated using a two-stage approach through body mass index) . We used RRs for SSB from the 2013 GBD as the RR values in the 2015 GBD did not have the two-stage approach calculated RRs directly available. A list of all included food-chronic disease combinations is presented in Table 1. In addition to the food-chronic disease combinations in Table 1, the GBD had also identified additional combinations (fruit: nasopharyngeal cancer, other pharyngeal cancer; sugar-sweetened beverages: gallbladder cancer, hypertensive heart disease, cardiomyopathy, atrial fibrillation, peripheral vascular, endocarditis, and other cardiovascular). However, we did not consider their costs as information on the costs of these diseases was not available through the EBIC . Because GBD age categories do not align exactly with EBIC age categories, we used the most conservative GBD relative risk estimates that applied to the relevant EBIC age category. A list of all relative risks used is presented in S1 Table.
Food consumption data analysis
Data on food consumption in the Canadian population were obtained from the Canadian Community Health Survey (CCHS) Cycle 2.2 (Nutrition) 2004, a cross-sectional national survey conducted by Statistics Canada and Health Canada from January 2004-January 2005  which at the time this study was conducted was the most recent national survey of dietary intakes of Canadians available since the Nutrition Canada survey conducted in 1970–1972. Briefly, this survey encompasses a representative sample of individuals residing in private dwellings from the ten Canadian provinces. This survey represents about 98% of individuals living in these provinces and the response rate was 76.5%.
The 24-hour dietary recall was administered using an automated multi-pass method . In total, 35,107 respondents completed a 24-hour dietary recall; of those respondents, 10,786 completed a second recall 3–10 days later. We eliminated the following records from all analyses: respondents <2 years, those marked as invalid, those where only breast milk was reported to be consumed, and records where no foods were recorded. In total, we included 44,325 24-hour dietary recall records that belonged to 33,932 respondents in our analyses. The Canadian Nutrient File (CNF 2001b) was used to determine the nutrient content of all reported foods.
Eating Well with Canada’s Food Guide (CFG)  was released in 2007 and outlines a dietary pattern for Canadians ≥2 years to help meet nutrient needs and prevent chronic diseases. Specific recommendations in terms of quantity and quality for four different food groups (Vegetables and Fruit, Grain Products, Milk and Alternatives, Meat and Alternatives) as well as other dietary factors (e.g., fats and oils, water) are outlined for nine different age and sex groups . Because all foods included in our analyses except sugar-sweetened beverages are part of various CFG groups, we quantified the consumption of these foods using CFG servings.
To quantify the number of CFG servings from the 24-hour dietary recalls, we used the Canada Food Guide file which classifies all included foods in every 24-hour dietary recall according to Health Canada’s validated Canadian Nutrient File/Canada Food Guide (CNF/CFG) tool [43,44]. This tool classifies foods into the four CFG groups as well as into subgroups within those groups (e.g., Vegetables and Fruit subgroup examples: a) fruit other than juice, b) vegetables dark green). In addition, within those subgroups, foods are categorized into tiers based on quantities of fat, sodium, and sugar and adjustments based on other food guide guidance. Foods in Tiers 1–3 are considered to count towards CFG servings, and foods belonging in Tiers 4 are higher in fat, sugars, and/or sodium and are generally not counted towards CFG servings . In addition, the CNF/CFG tool classifies some other foods (e.g., beverages).
A list of all food groups and subgroups included in our analyses are presented in Table 1. We included all foods belonging to each of the relevant subgroups with a few exceptions. For red meat, we only included foods from the a) beef, game and organ meats, and b) other meats (pork, veal, lamb) subgroups that met the International Agency for Research on Cancer red meat definition (i.e., meat from the muscle of mammals (e.g., pork, beef, veal, lamb)) . We also excluded fortified soy beverages from the fluid milk and fortified soy-based beverages subgroup, and potatoes and corn from relevant vegetable subgroups . For healthful foods (fruit, vegetables, whole grains, nuts and seeds, milk), we only included foods belonging to Tiers 1–3. For harmful foods (processed meat, red meat), we included foods belonging to Tiers 1–4. For beverages part of the CNF/CFG beverage subgroups, we considered those with ≥50kcal/226.8g to be a SSB . We assumed standard CFG serving sizes for each included food, and that one serving of SSB was 226.8g (Table 1).
We assumed that meeting food recommendations in Canada set by government and other reputable organizations were associated with the lowest chronic disease risk. Canada’s Food Guide recommendations were used for whole grains and milk. Because CFG does not contain separate vegetable and fruit recommendations (only a recommendation for vegetables and fruit combined), we estimated these recommendations using the GBD 2015 Theoretical Minimum Risk Exposure (TMRE). We first added the TMRE values for vegetables and fruit (~800g/day)  and determined the fraction of this total for vegetables (~0.63) and fruit (~0.38). For each of the nine CFG age and sex groups, we then multiplied the total number of CFG servings for vegetables and fruit by these fractions, and rounded to the nearest serving to estimate separate vegetable and fruit serving recommendations. Because CFG age groups do not exactly align with age groups used in the EBIC to estimate economic costs, we made the following assumptions. First, for the EBIC age category encompassing individuals ≤14 years, we used the CFG recommendations for children 4–8 years as these recommendations were in the middle of this age category. For adults, we used the CFG recommendation that applied to the majority of the relevant EBIC age and sex group. For nuts and seeds, processed meat, and red meat, we used recommendations set by reputable Canadian organizations because CFG does not provide specific recommendations for these foods. We used a recommendation of 30g/day of nuts and seeds (~1 CFG serving of nuts and seeds) based on the 2016 Canadian Cardiovascular Society Guidelines for the Management of Dyslipidemia for the Prevention of Cardiovascular Disease in the Adult . For processed meat and red meat, we used Canadian Cancer Society recommendations (processed meat: only for special occasions (assumed <0.05 servings/day based on the 2015 GBD TMRE); red meat: ≤3X85g cooked servings/week) . We further assumed a recommendation of 5g/day of SSB based on the 2015 GBD TMRE. A summary of all recommendations we used is provided in Table 1.
We estimated usual intake distributions (distribution of long-term average daily intakes) for each age and sex group of interest using the National Cancer Institute (NCI) method  MIXTRAN and DISTRIB macros executed with SAS version 9.4 (SAS Inc. Cary, NC). The NCI method has the capability to estimate usual intake distributions of dietary factors consumed either non-episodically (i.e., by almost everyone), using a one-part amount model, and episodically (i.e., those not consumed daily by nearly everyone), using a two-part amount and probability model [49,50]. This method also allows for the adjustment of covariates. Similar to other studies using the NCI method to calculate usual intake distributions [51–53], we stratified the sample into three groups: children ≤14 years, women ≥15 years, and men ≥15 years. Also similar to previous studies [51–53], we used dummy variables for sex (child models only) and age groups (15–34 years; 35–54 years; 55–64 years; 65–74 years; 75+ years) (adult models only), weekend or weekday recall, and sequence (difference between first and second recall). We used the two-part model for all foods except vegetables (all models), and milk (child model only) where we used a one-part model. We used one-day intakes for nuts and seeds because so few respondents consumed this food.
Using the DISTRIB macro, we obtained information on the percentage of each EBIC age and sex group (males and females; ≤14 years, 15–34 years, 35–54 years, 55–64 years, 65–74 years, ≥75 years) consuming different levels of food intakes. For all foods except SSBs, we obtained information on proportions of the population consuming half CFG serving increments. For sugar-sweetened beverages, we obtained information on the proportion of the population consuming increments of 226.8g (one serving) with the exception of capturing information on half serving increments between one and two servings. For foods where the recommendation is near 0, we assumed that the proportion of the population consuming below the following cut-points was not associated with elevated risk: 0.25 servings/day of processed meat and 56.7g/day of SSB as accurate estimates below these values could not be obtained.
Population attributable fraction calculations
We used the dose-response relative risks and food consumption data as inputs to calculate PAFs. The PAF uses both the relative risk of disease and the risk factor prevalence to estimate the fraction of disease that could theoretically be eliminated should the entire population follow healthy eating recommendations . The standard PAF equation is as follows: P(RR-1)/[P(RR-1)+1] where P is the risk factor prevalence, and RR is the relative risk of disease. However, because food intakes are associated with different levels of disease risk depending on the degree to which recommendations are met, we used the method outlined by Krueger et al  to handle multiple risk exposure levels. The equation is as follows: Where,
Pi is the proportion of individuals in interval i,
i (interval) refers to the consumption of specific numbers of servings/day (e.g., <0.5, 0.5-<1, ≥1),
RR is the relative risk for each incremental increase in food consumption,
is the relative risk for interval i relative to the recommended number of servings,
Xi is the mid value of interval i,
L is the recommended number of servings, and
n is the number of intervals above or below the recommended number of servings.
We calculated separate PAFs for each food for each chronic disease by sex for each of the following age groups: ≤14 years, 15–34 years, 35–54 years, 55–64 years, 65–74 years, and ≥75 years. The PAF values for each chronic disease were then combined for each sex and age group. Because some of the risk factors may be overlapping (i.e., several risk factors may be similar to one another and therefore redundant) and it is unclear what this overlap entails, it is important that we do not combine PAF values by adding them together as this would result in double counting. As suggested by Krueger et al , we used the following equation for this process: Combined PAF = 1-[(1-PAF1)(1-PAF2)(1-PAFn)]. This formula makes the assumption that the risk factors are independent (presence of one risk factor does not affect the probability of having another) and the joint effects of these risk factors are multiplicative (i.e. the presence of one factor does not affect the effect of another on a multiplicative scale). This approach has also been previously used to combine dietary risk factors for other outcomes (e.g., mortality) [37,54,55].
Estimation of direct health care and indirect costs
The calculated PAF outcomes were used to estimate the direct health care and indirect costs attributable to unhealthy eating. Multiple different sources released in different years were used to estimate costs. First, the proportion of direct health care costs (hospital, physician, drug) associated with each relevant chronic disease by sex for different age groups (i.e., ≤14 years, 15–34 years, 35–54 years, 55–64 years, 65–74 years, and ≥75 years) were extracted from the 2008 EBIC  which contains the most recent estimates of this information. Because this resource does not provide separate costs for different types of diabetes, we assumed that type 2 diabetes was responsible for 0.96 of total diabetes costs . Second, these proportions were then multiplied by the total hospital, physician, and drug costs reported by the 2014 National Health Expenditure Trends (i.e., 2014 costs in Canadian dollars)  to estimate more current direct health care costs for each relevant chronic disease for each sex and age group. The National Health Expenditure Trends resource contains information on the total hospital, physician, and drug costs for all diseases, age, and sex groups combined. Indirect costs were then estimated using the modified human capital approach as described by Krueger et al . Briefly, using the EBIC 1998  resource, a ratio of total direct health care to indirect costs (which includes costs associated with mortality, long-term disability, and short-term disability) was calculated for each included disease; these ratios are presented in Table 2. The indirect cost estimates in EBIC include discounted (5%) present value of lost productivity of all deaths during their estimated life expectancy, together with annual lost productivity by long term and short term disability. The used method utilizes age- and sex-specific rates of life expectancy, average annual earnings, workforce participation rates and values of unpaid work in Canadian provinces and territories . The 2014 hospital, physician, and drug costs were then multiplied by this ratio to obtain indirect costs for each disease for each age and sex group. Third, the 2014 direct health care (hospital, physician, drug) and indirect costs were then multiplied by the relevant PAF to determine the costs attributable to unhealthy eating. Similar to Krueger et al [6,7], a disaggregation step was also applied to determine the economic burden attributable to individual foods.
A sensitivity analysis was also conducted similar to Krueger et al . This analysis was done by re-calculating all PAFs using the 95% confidence interval upper and lower boundary estimates for all included relative risks. We also re-calculated the PAFs with the assumption that all individuals consumed one half serving (or one serving for SSBs) closer to the recommendations to determine the potential annual cost difference.
Table 3 shows the percentage of the Canadian population ≥2 years by age group and sex who met each food recommendation. Food recommendations that were least often met were for nuts and seeds (range: 0.8%-4.3%), and for whole grains (range: 1.3%-9.1%). Compared to nuts and seeds and whole grains, more Canadians met recommendations for fruit (range: 5.2%-19.0%), vegetables (range: 5.8%-26.1%), and milk (range: 4.7%-37.8%). The percentage of the population consuming <56.7g of SSB/day ranged from 9.9% to 68.9%, the percentage consuming <0.25 servings/day of processed meat ranged from 22.4% to 84.0%, and the percentage consuming less than ≤0.5 servings of red meat/day ranged from 17.1% to 70.1%. On average, Canadians were more likely to meet recommendations for harmful foods than those of healthful foods. More detailed information about the percentage of the population consuming specific numbers of food servings by age group and sex is presented in S2 Table.
Combined PAF values for each disease by age and sex are presented in Table 4. Combined PAF values for males were generally higher compared to females. Combined PAF values for individuals ≥35 years for cardiovascular diseases and type 2 diabetes were higher (PAF range: 13.6%-74.6%) than those for various cancers and chronic kidney disease (PAF range: 0.03%-27.6%).
Not meeting recommendations for the eight foods that have established recommendations was estimated to be responsible for an economic burden of CAD$13.8 billion (CAD$5.1 billion in direct health care costs, CAD$8.7 billion in indirect costs). The estimated direct health care costs that were attributable to not meeting these food recommendations represents approximately 3.9% of all hospital, physician, and drug costs in Canada in 2014. About 1/3 of the estimated costs were attributable to females and 2/3 were attributable to males (females: CAD$1.7 billion/year direct health care costs; CAD$2.8 billion indirect costs; males: CAD$3.5 billion/year direct health care costs, CAD$5.8 billion/year indirect costs). Over 50% of these estimated direct health care and indirect costs were attributable to ischemic heart disease (CAD$2.9 billion direct health care costs; CAD$4.9 billion indirect costs); type 2 diabetes was also a substantial contributor (CAD$1.7 billion direct health care costs; CAD$2.1 billion in indirect costs). Fewer of these estimated costs were attributable to cancer, stroke, and chronic kidney disease (cancer: CAD$253.9 million direct health care costs, CAD$1.2 billion in indirect costs; stroke: CAD$281.1 million in direct health care costs, CAD$480.4 in indirect costs; chronic kidney disease: CAD$4.4 million in direct health care costs, CAD$1.6 million in indirect costs). Our sensitivity analysis that used the relative risk 95% confidence interval boundaries in the PAF calculations revealed 95% confidence interval for total estimated costs to be $6.9-$18.5 billion. For estimated direct health care costs this was $2.6-$6.8 billion, and for estimated indirect costs this was $4.3-$11.7 billion.
Table 5 shows the estimated economic burden for each of the eight foods that have established recommendations. The estimated economic burden of inadequate intakes of vegetables, fruit, whole grains, nuts and seeds, and milk exceeds the burden of excess intakes of red meat, processed meat, and sugar-sweetened beverages (CAD$10.6 billion vs. CAD$3.2 billion). Overall, >20% of estimated direct health care and indirect costs attributable to unhealthy eating were due to each of inadequate nuts and seeds and whole grains with both having $1.3 billion direct health care costs and $2.0 billion indirect costs. Excess processed meat and inadequate intakes of fruit were estimated to each be responsible for about 14–17% of direct health care and indirect costs attributable to unhealthy eating (fruit: $780.9 million direct heath care costs, $1.4 billion indirect costs; processed meat: $728.3 million direct health care costs, $1.2 billion indirect costs). Excess intakes of sugar-sweetened beverages and inadequate intakes of vegetables were estimated each to be responsible for 6–9% of these costs (sugar-sweetened beverages: $382.8 million direct health care costs, $480.4 million indirect costs; vegetables: $430.4 million direct health care costs, $735.7 million indirect costs). Milk and red meat were estimated to be responsible for lowest economic burden at ≤6% (red meat: $134.1 million direct health care costs, $263.4 million in indirect costs; milk: $112.3 million direct health care costs; $536.4 million indirect costs).
We also calculated the estimated cost savings if Canadians who do not meet recommendations for any of the eight foods improved their dietary choices by consuming half a serving closer to the recommendations for fruit, vegetables, milk, whole grains, nuts and seeds, red meat, and processed meat, and everyone who did not meet recommendations for sugar-sweetened beverages consumed one serving closer to the recommendations. In this scenario, the estimated economic burden would be $8.9 billion (direct health care: $3.3 billion; indirect: $5.5 billion), which is $4.9 billion (direct health care: $1.8 billion; indirect: $3.1 billion) less relative to the burden without this assumption. Because severe allergies to nuts are a concern in Canada , we repeated this calculation without any changes with respect to nuts and seeds. In this scenario, the estimated economic burden was estimated to be $10.8 billion (direct health care: $4.1 billion; indirect: $6.8 billion).
We estimated that CAD$13.8 billion in direct health care and indirect costs (which could be as low as CAD$6.9 billion and as high as CAD$18.5 billion) in 2014 were associated with not meeting Canadian food recommendations. Our analysis revealed that the largest contributors to this economic burden were inadequate intakes of nuts and seeds and whole grains. These estimates are insightful for public health professionals when considering the range of potential policy options and programs that aim to reduce the health and economic burden of chronic diseases.
Our estimate of CAD$13.8 billion for the economic burden of not meeting food recommendations suggests that investments in promotion of healthy eating have the potential of substantial savings in direct health care and indirect costs in Canada. Krueger et al  estimated that the economic burdens attributable to smoking, alcohol consumption, and physical inactivity in Canada were CAD$19.5 billion, CAD$10.6 billion, and CAD$9.3 billion, respectively, in 2013. Similar to our approach, theirs involved use of population attributable fractions and comparable methods to estimate costs. However, it is important to note differences between our estimates and those by Krueger et al . Our focus was strictly on costs associated with the treatment and management of these chronic diseases. We had therefore only considered physician, hospital care and drug costs for these chronic diseases in our estimations. Krueger et al  took a broader perspective to costs to health care and therefore also considered costs for other health professionals (excluding dental), other health care expenditures, and health research. If we had considered the broader range of costs used by Krueger et al  our estimates would have been 30% to 50% higher. Adding 30% or 50% to our estimate of CAD$13.8 billion (increase to $17.9 billion or $20.7 billion) would put the economic burden of unhealthy eating at the same level than smoking and above physical inactivity estimated by Krueger et al . The economic burden of unhealthy eating combined with that of other modifiable behavioral risk factors (i.e., smoking, alcohol consumption, physical inactivity) would be estimated to cost Canadian society in excess of $50 billion/year . A status quo in promotion of these health behaviors will cause this enormous burden to continue. We refer to this as the ‘costs of doing nothing’ but recommend action.
Several previous studies conducted both in Canada [13,24,31,32] and internationally [4,25,26,28–30,34] have provided estimates for the economic burden of unhealthy eating. These studies have used various approaches and methodologies (e.g., top-down vs. bottom-up approaches; inclusion of varied types and numbers of dietary factors and types of costs) and have been conducted in different countries (each with different health care systems, chronic disease prevalence rates, and population distributions), which makes comparisons across studies difficult. Our study is unique as we used a bottom-up approach that considered all foods for which established recommendations exit, considered all chronic diseases for which scientific evidence exit, and considered age and sex sub-groups to add precision to the estimates. The only other estimate for Canada that was not limited to a single food item had followed a top down approach, i.e. did not consider specific food recommendations and dietary intakes. This study estimated the economic burden 20 years ago to be $6.6 billion ($1.3 billion in direct health care costs) .
The estimated economic burden associated with the consumption of the eight different foods are the result of a combination of their relative risk of disease, and the percentage of the population not meeting recommendations. These estimates help to draw attention to dietary intakes that are associated with the greatest burden of disease (e.g., nuts and seeds; whole grains). These estimates also illustrate that relatively small positive changes in dietary choices at the population level can lead to substantial cost savings. However, our estimates do not provide decision makers with actionable solutions on how to decrease the economic burden of disease. Although the estimated costs of inadequate nuts and seeds and whole grains exceeds that of vegetables and fruit, decision makers will also have to consider the possibility that it may be more difficult to increase intakes of nuts and seeds and whole grains because fewer Canadians meet those intakes compared to vegetables and fruit. The estimates from this study are useful to help guide the direction of future intervention and simulation studies examining the cost-effectiveness and cost-benefits of various programs and policy options (including costs to the consumer of eating more and specific types of healthy foods) to improve dietary intake. These studies, once completed, will provide guidance on how to best move forward with programs and policies to best decrease the burden of chronic diseases associated with unhealthy eating.
This study has several strengths. Dietary intakes were assessed using 24-hour dietary recalls of a representative sample of 33,932 Canadians. As well, only foods that cause chronic disease as identified by the GBD were included. A limitation is that our calculated value is likely an underestimate as the costs of certain chronic diseases and also costs associated with nutrients (e.g., fibre, fat, sodium) were not included. In addition, dietary assessment is a difficult process and underreporting unhealthy foods and over reporting of healthy foods is a common phenomenon. However, even with this limitation, the estimated costs associated with not meeting food recommendations were substantial. We also assumed that the proportion of costs associated with different diseases and ratio of direct health care to indirect costs remained consistent over time which may not be the case; however, these assumptions have been used in other similar Canadian studies [5,7,13,24]. We also did not examine regional differences in costs which have been previously reported to vary for other behavioral chronic disease risk factors in Canada (e.g., smoking, excess body weight, and physical inactivity ). In addition, these results do not show the estimated lifetime costs of not meeting food-based recommendations in Canada which is a limitation in all prevalence-based cost-of-illness studies. Using an incidence-based approach which is more complex would allow the lifetime costs to be estimated. A further limitation is that the dietary intake data used for this study was collected in 2004; however, this was the most recent data that is available with information on dietary intakes from a representative sample of Canadians at the time this study was conducted. We encourage researchers to use national nutrition datasets in their jurisdictions to estimate the economic burden of not meeting their nation’s food recommendations using a similar approach.
The economic burden of not meeting food recommendations in Canada was estimated to be $13.8 billion in 2014. This estimate allows the impact of our food choices to be compared to estimates associated with other behavioral lifestyle choices (e.g., smoking, physical inactivity) determined using similar approaches. Although making dietary changes is a complex process which requires strong support and major investments from government and society, policies and programs which have the ability to make small positive changes at the population level have the potential to reduce the currently tremendous economic burden.
We would like to thank Irene Wong (University of Alberta Research Data Centre Analyst) for her assistance in accessing the Canadian Community Health Survey data.
- 1. Statistics Canada. Leading causes of death, by sex. Available at: http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/hlth36a-eng.htm Accessed 22 Dec 2017.
- 2. Public Health Agency of Canada. Economic Burden of Illness in Canada, 2005–2008. Ottawa: Public Health Agency of Canada, 2014.
- 3. World Health Organization. 2008–2013 Action Plan for the Global Strategy for the Prevention and Control of Noncommunicable Diseases. Available at: http://whqlibdoc.who.int/publications/2009/9789241597418_eng.pdf Accessed 22 Dec 2017.
- 4. Scarborough P, Bhatnagar P, Wickramasinghe KK, Allender S, Foster C, Rayner M. The economic burden of ill health due to diet, physical inactivity, smoking, alcohol and obesity in the UK: An update to 2006–07 NHS costs. J Public Health (Bangkok). 2011;33: 527–535.
- 5. Krueger H, Krueger J, Koot J. Variation across Canada in the economic burden attributable to excess weight, tobacco smoking and physical inactivity. Can J Public Heal. 2015;106: e171–e177.
- 6. Krueger H, Williams D, Ready AE, Trenaman L, Turner D. Improved estimation of the health and economic burden of chronic disease risk factors in Manitoba. Chronic Dis Inj Can. 2013;33: 236–46. pmid:23987220
- 7. Krueger H, Andres EN, Koot JM, Reilly BD. The economic burden of cancers attributable to tobacco smoking, excess weight, alcohol use, and physical inactivity in Canada. Curr Oncol. 2016;23: 241–249. pmid:27536174
- 8. Kaiserman MJ. The cost of smoking in Canada, 1991. Chronic Dis Can. 1997;18: 13–19. pmid:9188515
- 9. Krueger H, Koot JM, Rasali DP, Gustin SE, Pennock M. Regional variations in the economic burden attributable to excess weight, physical inactivity and tobacco smoking across British Columbia. Heal Promot Chronic Dis Prev Canada 2016;36: 76–86.
- 10. Choi BC, Pak AW. Health and social costs of tobacco use in Ontario, Canada, 1979 and 1988. J Epidemiol Community Health. 1996;50: 81–85. pmid:8762360
- 11. Xie X, Robson L, Single E, Rehm J, Paul J. The economic consequences of smoking in Ontario. Pharmacol Res. 1999;39: 185–191. pmid:10094843
- 12. Rehm J, Gnam W, Popova S, Baliunas D, Brochu S, Fischer B, et al. The costs of alcohol, illegal drugs, and tobacco in Canada, 2002. J Stud Alcohol Drugs. 2007;68: 886–895. pmid:17960307
- 13. Krueger H, Koot J, Andres E. The economic benefits of fruit and vegetable consumption in Canada. Can J Public Heal. 2017;102: e152–e161.
- 14. Katzmarzyk PT, Janssen I. The economic costs associated with physical inactivity and obesity in Canada: an update. Can J Appl Physiol. 2004;29: 90–115. pmid:15001807
- 15. Birmingham CL, Muller JL, Palepu A, Spinelli JJ, Anis AH. The cost of obesity in Canada. CMAJ 1999;160: 483–488. pmid:10081464
- 16. Anis AH, Zhang W, Bansback N, Guh DP, Amarsi Z, Birmingham CL. Obesity and overweight in Canada: An updated cost-of-illness study. Obes Rev. 2010;11: 31–40. pmid:19413707
- 17. Tran BX, Nair AV, Kuhle S, Ohinmaa A, Veugelers PJ. Cost analyses of obesity in Canada: scope, quality, and implications. Cost Eff Resour Alloc. 2013;11: 3. pmid:23394349
- 18. Moffatt E, Shack LG, Petz GJ, Sauvé JK, Hayward K, Colman R. The cost of obesity and overweight in 2005: A case study of Alberta, Canada. Can J Public Heal. 2011;102: 144–148.
- 19. Blouin C, Hamel D, Vandal N, Barry AD, Lo E, Lacroix G, et al. The economic consequences of obesity and overweight among adults in Quebec. Can J Public Heal. 2016;10: 507–513.
- 20. Katzmarzyk PT, Gledhill N & Shephard RJ. The economic burden of physical inactivity in Canada. CMAJ. 2000;163: 1435–1440. pmid:11192648
- 21. Janssen I. Health care costs of physical inactivity in Canadian adults. Appl Physiol Nutr Metab. 2012;37: 803–806. pmid:22667697
- 22. Taylor G. The Chief Public Health Officer’s Report on the State of Public Health in Canada 2015: Alcohol Consumption in Canada. Ottawa: Public Health Agency of Canada, 2016.
- 23. Garriguet D. Canadians’ eating habits. Health Rep 2007;18: 17–32. pmid:17578013
- 24. Ekwaru JP, Ohinmaa A, Loehr S, Setayeshgar S, Thanh NX, Veugelers PJ. The economic burden of inadequate consumption of vegetables and fruit in Canada. Public Health Nutr. 2016;20: 515–523. pmid:27819197
- 25. Barnard ND, Nicholson A, Howard JL. The medical costs attributable to meat consumption. Prev Med. 1995;24: 646–655. pmid:8610089
- 26. Cadilhac DA, Magnus A, Sheppard L, Cumming TB, Pearce DC, Carter R. The societal benefits of reducing six behavioural risk factors: an economic modelling study from Australia. BMC Public Health 2011;11: 483. pmid:21689461
- 27. Manuel DG, Perez R, Benne C, Laporte A, Wilton AS, Gandhi S, et al. A $4.9 Billion Decrease in Health Care Expenditure: The Ten-Year Impact of Changing Smoking, Alcohol, Diet and Physical Activity on Health Care Use in Ontario. Toronto: Institute for Clinical Evaluative Sciences, 2016.
- 28. Doidge JC, Segal L, Gospodarevskaya E. Attributable risk analysis reveals potential healthcare savings from increased consumption of dairy products. J Nutr. 2012;142: 1772–1780. pmid:22833660
- 29. McCarron DA, Heaney RP. Estimated healthcare savings associated with adequate dairy food intake. Am J Hypertens. 2004;17: 88–97. pmid:14700520
- 30. Meier T, Deumelandt P, Christen O, Stangl GI, Riedel K, Langer M. Global burden of sugar-related dental diseases in 168 countries and corresponding health care costs. J Dent Res. 2017;96: 845–854. pmid:28489474
- 31. Jones AC, Veerman JL, Hammond D. The Health and Economic Impact of a Tax on Sugary Drinks in Canada. March 2017.
- 32. Abdullah MM, Gyles CL, Marinangeli CP, Carlberg JG, Jones PJ. Cost-of-illness analysis reveals potential healthcare savings with reductions in type 2 diabetes and cardiovascular disease following recommended intakes of dietary fiber in Canada. Front Pharmacol. 2015;6: 167. pmid:26321953
- 33. Cancer Care Ontario, Ontario Agency for Health Protection and Promotion (Public Health Ontario). Taking action to prevent chronic disease: recommendations for a healthier Ontario. Toronto: Queen’s Printer for Ontario, 2012.
- 34. Frazão E. High Costs of Poor Eating Patterns in the United States. In: Frazão E, ed. America’s Eating Habits: Changes and Consequences. Agriculture Information Bulletin No. AIB750. Washington, DC: U.S. Department of Agriculture, Economic Research Service, Food and Rural Economics Division; 1999:5–32.
- 35. Healthy Aging and Wellness Working Group. Healthy aging in Canada: A new vision, a vital investment from evidence to action. A background paper prepared for the federal, provincial and territorial committee of officials (seniors). 2006. Available at: http://www.health.gov.bc.ca/library/publications/year/2006/Healthy_Aging_A_Vital_latest_copy_October_2006.pdf Accessed 22 Dec 2017.
- 36. Health Canada (2003) Economic Research Analysis Section, Policy Research Division, Strategic Policy Directorate, Population and Public Health Branch, Health Canada. Custom Tabulations.
- 37. GBD 2015 Risk Factors Collaborators. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet 2016;388: 1659–1724. pmid:27733284
- 38. World Cancer Research Fund / American Institute for Cancer Research. Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective. Washington DC: AICR, 2007.
- 39. GBD 2013 Risk Factors Collaborators. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2015;386: 1990–2013.
- 40. Health Canada. Canadian Community Health Survey, Cycle 2.2, Nutrition (2004)—A Guide to Accessing and Interpreting the Data. Ottawa: Health Canada; 2006.
- 41. United States Department of Agriculture. AMPM—USDA Automated Multiple-Pass Method: USDA ARS. Available at: https://www.ars.usda.gov/northeast-area/beltsville-md/beltsville-human-nutrition-research-center/food-surveys-research-group/docs/ampm-usda-automated-multiple-pass-method/ Accessed 22 Dec 2017.
- 42. Health Canada. Eating Well with Canada's Food Guide. Ottawa: Health Canada, 2007.
- 43. Health Canada. The Development and Use of a Surveillance Tool: The Classification of Foods in the Canadian Nutrient File According to Eating Well with Canada’s Food Guide. Ottawa Health Canada, 2014.
- 44. Munene LAE, Dumais L, Esslinger K, Jones-Mclean E, Mansfield E, Verreault MF, et al. A surveillance tool to assess diets according to Eating Well with Canada’s Food Guide. Heal Reports 2015;26: 12–20.
- 45. International Agency for Research on Cancer. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Available at: http://monographs.iarc.fr/index.php. Accessed 5 Nov 2016.
- 46. Micha R, Kalantarian S, Wirojratana P, Byers T, Danaei G, Elmadfa I, et al. Estimating the global and regional burden of suboptimal nutrition on chronic disease: methods and inputs to the analysis. Eur J Clin Nutr. 2012;66: 119–129. pmid:21915137
- 47. Anderson TJ, Grégoire J, Pearson GJ, Barry AR, Couture P, Dawes M, et al. 2016 Canadian Cardiovascular Society Guidelines for the Management of Dyslipidemia for the Prevention of Cardiovascular Disease in the Adult. Can J Cardiol. 2016;32: 1263–1282. pmid:27712954
- 48. Canadian Cancer Society. Red and processed meat. Available at: http://www.cancer.ca/en/prevention-and-screening/live-well/nutrition-and-fitness/eating-well/red-and-processed-meat/?region=on. Accessed: 2 May 2017.
- 49. Tooze JA, Midthune D, Dodd KW, Freedman LS, Krebs-Smith SM, Subar AF, et al. A new statistical method for estimating the usual intake of episodically consumed foods with application to their distribution. J Am Diet Assoc. 2006;106: 1575–1587. pmid:17000190
- 50. Tooze JA, Kipnis V, Buckman DW, Carroll RJ, Freedman LS, Guenther PM, et al. A mixed-effects model approach for estimating the distribution of usual intake of nutrients: the NCI method. Stat Med. 2010;29: 2857–2868. pmid:20862656
- 51. Kirkpatrick SI, Dodd KW, Parsons R, Ng C, Garriguet D, Tarasuk V. Household food insecurity is a stronger marker of adequacy of nutrient intakes among Canadian compared to American youth and adults. J Nutr. 2015;145: 1596–1603. pmid:25995277
- 52. Krebs-Smith SM, Guenther PM, Subar AF, Kirkpatrick SI, Dodd KW. Americans do not meet federal dietary recommendations. J Nutr. 2010;140: 1832–1838. pmid:20702750
- 53. Kirkpatrick SI, Dodd KW, Reedy J, Krebs-Smith SM. Income and race/ethnicity are associated with adherence to food-based dietary guidance among US adults and children. J Acad Nutr Diet. 2012;112: 624–635.e6. pmid:22709767
- 54. Azevedo ESG, De Moura L, Curado MP, Gomes Fda S, Otero U, Rezende LF, et al. The fraction of cancer attributable to ways of life, infections, occupation, and environmental agents in Brazil in 2020. PLoS One. 2016;11: 1–13.
- 55. Whiteman DC, Webb PM, Green AC, Neale RE, Fritschi L, Bain CJ, et al. Cancers in Australia in 2010 attributable to modifiable factors: Introduction and overview. Aust N Z J Public Health. 2015;39: 403–407. pmid:26437722
- 56. Johnson JA, Pohar SL, Majumdar SR. Health care use and costs in the decade after identification of type 1 and type 2 diabetes. Diabetes Care 2006;29: 2403–2408. pmid:17065675
- 57. Canadian Institute for Health Information. National Health Expenditure Trends, 1975 to 2015. Ottawa: CIHI, 2015.
- 58. Health Canada. Economic Burden of Illness in Canada, 1998. Ottawa: Health Canada, 2002.
- 59. Krueger H, Turner D, Krueger J, Ready AE. The economic benefits of risk factor reduction in Canada: Tobacco smoking, excess weight and physical inactivity. Can J Public Heal. 2014;105: 69–78.
- 60. Soller L, Ben-Shoshan M, Harrington DW, Knoll M, Fragapane J, Joseph L, et al. Adjusting for nonresponse bias corrects overestimates of food allergy prevalence. J Allergy Clin Immunol Pract. 2015;3:291–293.e2. pmid:25609350