Cardiovascular disease (CVD) is a significant health burden globally and while the age-standardized prevalence of CVD has declined in the US and other high-income countries over the last decades, it remains a leading cause of morbidity and mortality [1
]. In 2016, prevalence of CVD, excluding hypertension, in the US population 20 years and older was 9.0%, representing 24.3 million adults; including hypertension, prevalence of CVD increases to 121.5 million adults [2
]. The economic burden associated with CVD is sizeable and medical costs alone are expected to exceed US$
750 billion by 2035 [2
Epidemiological and clinical evidence provide support for the importance of diet in maintaining health and reducing risk for chronic disease. For example, consumption of whole grains is associated with several health benefits, including reduced risk for CVD [3
]. The 2015–2020 Dietary Guidelines for Americans (DGA) recommends consumption of at least half of total grains as whole grains [5
]. For a 2000-calorie diet, the recommended amount of grains in the Healthy U.S.-Style Eating Pattern is 6 ounce-equivalents per day (oz-eq/day) with at least 3 oz-eq/day consumed as whole grains. On a given day, however, whole grains account for approximately 16% of total grains consumed among US adults [6
], indicating that most adults fall far short of meeting the dietary recommendation.
Previous scenario analyses have modeled changes in healthcare costs from beneficial shifts in dietary patterns and provide an indication of the potentially substantial economic impact associated with modifications to the population’s diet [7
]. The purpose of this study is to estimate the impact on healthcare costs associated with a reduced incidence of total CVD based on modeling incremental increases in whole grains consumption as a proportion of total grains up to levels recommended in the DGA.
Using risk reduction parameters derived from the literature (Table 1
), the estimated annual direct medical cost savings from reduced risk of CVD was US$
21.9B (range of US$
5.5B to US$
38.4B) assuming half of all grains consumed are whole grains (i.e., an increase of 2.24 oz-eq/day), as shown in Table 2
. The estimated annual direct medical cost savings from reduced risk of CHD was US$
8.4B to US$
22.4B) assuming half of all grains consumed are whole grains. Proportionally lower cost savings were estimated in models assuming smaller increases in whole grains consumption, with increased intake of 0.25 oz-eq/day estimated to save US$
2.4B annually (US$
0.6B to US$
4.3B) from reduced risk of CVD.
In the sensitivity analysis, total costs (direct medical and indirect) were applied to estimates of reduction in risk of incidence or mortality (Table 3
). Under these analyses, the estimated annual cost savings with an increase of 2.24 oz-eq/day in whole grains consumption was US$
9.0B to US$
63.1B) from reduced risk of CVD and US$
16.8B to US$
44.9B) from reduced risk of CHD.
Findings from this study indicate that increased consumption of whole grains as a proportion of total grains to align with dietary recommendations may result in substantial healthcare cost savings due to reduced risk for cardiovascular diseases. More complex approaches to health economic modeling are available such as cohort models and microsimulation models [16
]. A cost-of-illness [7
] or cost-benefit [10
] approach, or a scenario analysis [11
] approach, similar to the current analysis are commonly used for public health nutrition models, and these approaches are particularly appropriate when there are limited data about consumption of a specific food, the lag between increased (or decreased) consumption and the health effect, and how other population risk factors may affect the change in incidence.
Loewen and colleagues estimated the economic burden (total direct health and indirect costs) of low whole grains consumption on cardiovascular outcomes including ischemic heart disease, ischemic stroke, and hemorrhagic stroke at CAD$
2.6 billion per year in 2018 [18
]. We are not aware of other models assessing healthcare savings from increased consumption of whole grains in the general population in the US, therefore the results from this analysis provide the first estimate of potential US cost savings when whole grain consumption is directly aligned with the DGA. Lee and colleagues’ model explored the cost-effectiveness of financial incentives associated with increased whole grains and other dietary improvements in Medicare and Medicaid-covered US beneficiaries [17
]. A 2019 microsimulation by Jardim and colleagues estimated that 9.3% of cardiometabolic direct medical costs in the US ($
7.5B) were attributable to suboptimal whole grain consumption [16
Given the substantial savings estimated with even a modest increase in whole grains consumption of 0.25 oz-eq/day, this study, along with others, supports the importance of promoting beneficial shifts in dietary patterns. These data may be particularly useful when establishing goals for public health programs, in that even dietary changes below DGA targets may have health economic benefits.
The change in the proportion of grains consumed as whole grains necessary to meet DGA recommendations is not trivial, and increasing consumption of whole grains may be a challenge. For example, an analysis of whole grains purchases by participants in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) following changes to the WIC package in 2009 show a monthly increase in whole grain purchases of 4.34 oz-eq, which corresponds to an increase from 28.4% to 34.5% of total grains as whole grains [19
]. While the increased purchases of whole grains reported are encouraging, these results indicate that substantial increases in whole grains consumption may be difficult; availability and access are necessary but not sufficient to encourage large changes in consumption. It is particularly challenging given that there is not yet consistency in how whole grains are defined. The Whole Grains Stamp is used globally to provide consumers with information on percent and total whole grain content [20
] and is useful in the absence of a labeling requirement. As the scientific community reaches consensus on definitions and communication mechanisms, it will enhance consumers’ ability to recognize specific amounts of whole grains, facilitating more informed replacement of refined grains with whole grains.
There are several important limitations to consider in interpreting findings as well as for designing future studies. First, there was no consensus in the literature on the definition of a whole grain food or the amount of whole grains that constitutes a ‘serving’, thus there are inconsistencies in how whole grains consumption is captured, a challenge that the research and manufacturing community is currently trying to solve. Second, the model assumes a linear relationship between increased consumption of whole grains and risk of CVD or CHD, where each unit increase in whole grains consumption proportionally reduces risk and in turn proportionally reduces cost. Evidence of non-linear associations between consumption of whole grains and risk of CVD and CHD was reported for the estimates used in this model [3
]. For CHD, the steepest dose-response was observed with consumption of up to 90 g whole grains (i.e., ~3 servings) while further, though less steep, reductions were seen with consumption of up to 210 g whole grains. For CVD, there was evidence of a linear dose-response up to approximately 50 g (i.e., ~1.5 servings) of whole grains and more modest declines with consumption of up to 200 g whole grains. Consequently, use of the effect size per 90 g to calculate the incremental change per unit serving of whole grains (30 g, assumed to be equivalent to 1 oz-eq) or fractions of a serving may underestimate reduction in CVD risk in the range of modeled changes. The assumption of linearity remains a concern, as is the potential that the association between incidence and costs is also not linear, which could further confound the results. The existing literature does not allow for the prediction of changes in the distribution of severity of disease; for example, there may not only be fewer adults with CVD, but there could be a shift toward milder and thus less costly disease. Similarly, the model assumes a single level of whole grains consumption based on the population level intakes; for some adults, a small change in whole grains consumption could meet DGA recommendations while for others following a Healthy US-Style dietary pattern, an increase of 2.24 oz-eq/day may be insufficient to meet recommendations. Analysis of whole grains intake by age, sex, and ethnicity suggests there are differences in consumption that might be applied to future cost analyses [6
]. As more of these relationships are documented and assumptions are replaced with evidence, microsimulation modeling on the impact of increases in whole grains consumption will become increasingly informative. The model in the current analysis does not address other potential health outcomes and related costs, either positive or negative, related to increased whole grains consumption in place of refined grains. The model also does not address potential cost to the consumer, though some data suggest a smaller price differential between whole grains and non-whole grains than between other healthier and traditional foods [21
]. As additional data are collected and identified to minimize the uncertainty in this preliminary analysis, this model can be refined.
Findings from this modeling study indicate that increasing whole grains consumption among US adults, from incremental increases to more substantial increases to align more closely with the DGA recommendation, has the potential to provide substantial savings in healthcare costs.