While food choice is influenced by a large number of factorsReference Shepherd1, the price of food is clearly an important determinantReference Glanz, Basil, Maibach, Goldberg and Snyder2, Reference Lennernas, Fjellstrom, Becker, Giachetti, Schmitt and Remaut de Winter3. An inverse relationship exists between the energy density of foods (energy derived from foods per edible weight of foods) and energy cost (monetary cost of foods per energy derived from foods)Reference Drewnowski and Darmon4. Generally, energy-dense and nutrient-dilute foods such as cereals, fats and oils, and sugar and sweets provide dietary energy at the lowest cost. Conversely, the cost per calorie of energy-dilute and nutrient-dense foods including fish and shellfish, vegetables and fruit is much higher.
If healthier foods cost more then so too will healthier diets, suggesting that consumers with limited financial resources might select energy-dense and nutrient-dilute diets as a means of saving money. Observational studies (albeit a limited number) on the cost of freely chosen diets have consistently shown that healthful diets are more expensive than less healthful dietsReference Stender, Skovby, Haraldsdottir, Andresen, Michaelsen and Nielsen5–Reference Schroder, Marrugat and Covas10. To our knowledge, however, all studies of self-selected diets and monetary costs have been conducted in European countries, with none reported in Asian countries, including Japan.
In contrast to Western populations, Japanese obtain the largest part of their energy intake from rice (29%)Reference Sasaki, Takahashi, Iitoi, Iwase, Kobayashi and Ishihara11. Further, fat intake is relatively low (≤30% energy)12. The relationship of dietary costs to dietary intake and diet quality may therefore differ between Western and Japanese populations. Here, we examined the monetary cost of dietary energy in relation to food and nutrient intake and energy density in a group of young Japanese women. Given recent findings from Spain of the higher monetary costs of healthy dietary patterns associated with a lower body mass index (BMI)Reference Schroder, Marrugat and Covas10, we also examined the association between monetary cost of dietary energy and BMI.
Subjects and methods
Subjects and survey procedure
The present study was based on a self-administered questionnaire survey among dietetic students (n = 4679) from 54 universities, colleges and technical schools in 33 of 47 prefectures in Japan. Staff at each institution distributed two questionnaires on dietary habits and other lifestyle items during the preceding month at orientation sessions or early lectures for freshman students who entered dietetic courses in April 2005, in most institutions within 2 weeks after the course began. Students filled out the questionnaires during the session, lecture or at home, and then submitted the completed forms to staff at each institution. The staff then checked the responses according to the survey protocol. When missing answers or logical errors were identified, the students were asked to complete the questionnaire again. The staff then mailed the completed questionnaires to the survey centre, where the answers were checked once more. Problematic questionnaires were returned to the staff at the respective institution, and the students were asked to complete the questionnaires again. All questionnaires were thus checked at least once each by staff at the respective institution and at the survey centre. Most surveys were completed by May 2005. The protocol of the study was approved by the Ethics Committee of the National Institute of Health and Nutrition.
In total, 4394 students (4168 women and 226 men) answered two questionnaires (response rate = 93.9%). For the present analysis, we selected female subjects aged 18–20 years (n = 4060). We then excluded those who were in an institution where the survey had been conducted at the end of May (n = 98), those with extremely low or high energy intake (<500 kcal day−1 or >4000 kcal day−1) (n = 23) and those with missing information on the variables used (n = 12). As some subjects were in more than one exclusion category, the final analysis sample comprised 3931 women.
Dietary intakes
Dietary habits during the preceding month were assessed using a previously validated, self-administered, diet history questionnaire (DHQ)Reference Sasaki, Yanagibori and Amano13–Reference Sasaki, Ushio, Amano, Morihara, Todoriki and Uehara15. This is a 16-page structured questionnaire that consists of the following seven sections: general dietary behaviour; major cooking methods; consumption frequency and amount of six alcoholic beverages; consumption frequency and semi-quantitative portion size of 122 selected food and non-alcoholic beverage items; dietary supplements; consumption frequency and semi-quantitative portion size of 19 cereals usually consumed as staple foods (rice, bread and noodles) and miso (fermented soybean paste) soup; and open-ended items for foods consumed regularly (once a week or more) but not appearing in the DHQ. Items and portion sizes were derived primarily from data in the National Nutrition Survey of Japan and several recipe books for Japanese dishesReference Sasaki, Yanagibori and Amano13.
Estimates of dietary intake for 148 food and beverage items, energy and nutrients were calculated using an ad hoc computer algorithm for the DHQ based on the Standard Tables of Food Composition in Japan 16. Information on dietary supplements and data from the open-ended questionnaire items were not used in the calculation. Energy-adjusted values of dietary intake were calculated using the percentage of energy for macronutrients and the amount per 1000 kcal for dietary fibre, vitamins, minerals and foods. Alcohol intake was not used because of an extremely low mean intake (0.8 g day−1). Dietary energy density (kcal g−1) was calculated by dividing total energy intake by the estimated edible weight of all foods and caloric beverages consumed (excluding alcohol). Detailed descriptions of the methods used to calculate dietary intake and the validity of the DHQ have been published elsewhereReference Sasaki, Yanagibori and Amano13–Reference Sasaki, Ushio, Amano, Morihara, Todoriki and Uehara15. Pearson correlation coefficients between the DHQ and 3-day estimated dietary records were 0.48 for energy, 0.48–0.55 for energy-yielding nutrients (excluding alcohol) and 0.19–0.68 for vitamins and minerals among 47 womenReference Sasaki, Yanagibori and Amano13. In addition, the Pearson correlation coefficients between the DHQ and 16-day weighed dietary records among 92 women were 0.69 for total dietary fibre, 0.40 for energy density and 0.33 for edible weight consumed, and the Spearman correlation coefficients for food groups ranged from 0.28 to 0.59 (Sasaki S, unpublished observations, 2004).
Dietary costs
Monetary costs of habitual diets obtained from the DHQ (Japanese yen day−1) were calculated by multiplying the amount of food consumed from the DHQ (g day−1) by the estimated price of each food (Japanese yen g−1) and summing the products (1 Japanese yen = 0.007 Euros = 0.008 US dollars in April 2006). The procedure for estimating costs was based on the assumption that all foods were purchased and then prepared and consumed at home. Alcoholic beverages (six items), non-alcoholic and non-caloric beverages (four items), drinking water, noodle soup, and water for miso soup were excluded from calculation. The price of each food was obtained mainly from the National Retail Price Survey 200417 (122 of 135 items; 90%). This survey is conducted annually in 167 villages, towns and cities, and average prices were calculated as mean values of all survey areas, weighted for population size. For food items whose prices are not published in the National Retail Price Survey (13 of 135 items; 10%), prices were taken from the websites of nationally distributed supermarket (Seiyu, Japan) and fast-food restaurant (McDonalds, Japan and Mister Donut, Japan) chains. Sale prices were not used to determine costs. Costs of combined foods such as pizza were calculated using prices of frozen equivalents. Calculation included correction for preparation and waste (e.g. trimming and peeling of vegetables and fruits, removal of bones and skin from fish). Monetary cost of dietary energy (Japanese yen 1000 kcal−1) was calculated by dividing the estimated daily cost of the diet (Japanese yen day−1) by the daily energy intake (kcal day−1) and multiplying by 1000. The monetary cost of dietary energy of each food is listed in Table 1.
Table 1 Monetary cost of dietary energy of each foodFootnote *
* Foods not listed above (six alcoholic beverages, four non-alcoholic and non-caloric beverages, drinking water, noodle soup, water for miso soup) were not used for the calculation of diet costs.
† 1 Japanese yen = 0.007 Euros = 0.008 US dollars in April 2006.
BMI
Body weight and height were self-reported as part of the DHQ. BMI was calculated as weight (kg) divided by the square of height (m).
Other variables
Residential areas, reported in the 12-page lifestyle questionnaire, were grouped into six regions (Hokkaido and Tohoku; Kanto; Hokuriku and Tokai; Kinki; Chugoku and Shikoku; Kyushu) based on blocks used in the National Nutrition Survey in Japan12 and hereafter referred to as ‘residential block’. The residential areas were also grouped into three categories according to population size (city with population ≥1 million; city with population <1 million; town and village) and hereafter referred to as ‘size of residential area’. The lifestyle questionnaire also assessed living status (living alone; living with family; living with others), current smoking (yes; no) and whether currently trying to lose weight (yes; no).
Subjects additionally reported in the questionnaire the time they usually got up and went to bed, which was used to calculate sleeping hours, and the frequency and duration of high- and moderate-intensity activities, walking and sedentary activities. Each activity was assigned a metabolic equivalent task (MET) value from a previously published table, namely 0.9 for sleeping, 7.0 for high-intensity activity, 5.0 for moderate-intensity activity, 3.3 for walking and 1.5 for sedentary activityReference Ainsworth, Haskell, Leon, Jacobs, Montoye and Sallis18, Reference Ainsworth, Haskell, Whitt, Irwin, Swartz and Strath19. The number of hours spent per day on each activity was multiplied by its MET value, and all MET-hour products were summed to give a total MET-hour score for the day, which essentially corresponded to the number of kilocalories per kilogram of body weight expended by the individual during the day. The standard value of basal metabolic rate for Japanese people was also expressed as the number of kilocalories per kilogram of body weight expended by an individual during the day. Physical activity level was then calculated by dividing the total MET-hour score by the standard value of basal metabolic rate for Japanese women aged 18–29 years20. Current alcohol drinking (yes; no), current dietary supplement use (yes; no) and rate of eating (very slow; relatively slow; medium; relatively fast; very fast) were assessed using the DHQ.
Underreporting of energy intake
Underreporting of energy intake is an ongoing controversy in studies using self-report instruments to collect dietary informationReference Goldberg, Black, Jebb, Cole, Murgatroyd and Coward21, Reference Black, Goldberg, Jebb, Livingstone, Cole and Prentice22. To estimate the prevalence of energy underreporting in this population, the ratio of reported energy intake to basal metabolic rate (estimated according to a published equation for Japanese women aged 18–29 years using reported body weight20) was computed. Using the Goldberg cut-off methodReference Goldberg, Black, Jebb, Cole, Murgatroyd and Coward21 recently re-evaluated by BlackReference Black23, a ratio of 1.09 was calculated as the lower cut-off point for reasonable habitual energy intake. Persons with ratios of energy intake to basal metabolic rate <1.09 were considered by this technique to be energy underreporters (n = 666; 17%). Results in all subjects and those obtained after exclusion of energy underreporters did not differ materially. Therefore, we only present the analyses of all subjects.
Statistical analysis
All statistical analyses were performed using SAS statistical software, version 8.2 (SAS Institute Inc.). We calculated both crude and multivariate-adjusted means of dietary intake and BMI by quintile categories of monetary cost of dietary energy. Confounding variables included in the multivariate models were residential block (Hokkaido and Tohoku; Kanto; Hokuriku and Tokai; Kinki; Chugoku and Shikoku; Kyushu), size of residential area (city with population ≥1 million; city with population <1 million; town and village), living status (living with family; living alone; living with others), current smoking (yes; no), current alcohol drinking (yes; no), current dietary supplement use (yes; no), currently trying to lose weight (yes; no), rate of eating (very slow; relatively slow; medium; relatively fast; very fast), and physical activity level (continuous) and energy intake (kcal day−1, continuous) for variables except for energy intake. For analyses on BMI, intakes of protein (% of energy, continuous), fat (% of energy, continuous) and dietary fibre (g 1000 kcal−1, continuous) were further included as confounding factors. Linear trends with increasing levels of monetary cost of dietary energy were tested by assigning to each participant the median value for the category and modelling this value as a continuous variable. All reported P-values are two-tailed and were considered statistically significant at the <0.05 level.
Results
Subject characteristics are shown in Table 2. Women in the higher quintiles of monetary cost of dietary energy tended to live in areas with larger populations, live with family, use alcohol and dietary supplements, be slow eaters and trying to lose weight. Women in the higher quintiles of monetary cost of dietary energy had higher mean values of body weight, BMI and physical activity level. Major contributors to total dietary cost were fish and shellfish (16%), meat (16%), vegetables (16%), confectioneries (12%) and rice (9%) (Table 3). Women in the higher quintiles of monetary cost of dietary energy had lower mean values of diet cost for cereals (except noodles). Mean costs of all other foods were higher in women in the higher quintiles of monetary cost of dietary energy.
Table 2 Subject characteristics according to quintile category of monetary cost of dietary energyFootnote *
* Values are mean ± standard deviation or number of subjects (%).
† 1 Japanese yen = 0.007 Euros = 0.008 US dollars in April 2006.
‡ For continuous variables, tests for linear trend used the median value in each quintile as a continuous variable in linear regression; a Mantel–Haenszel χ2 test was used for categorical variables.
Table 3 Total dietary cost according to quintile category of monetary cost of dietary energyFootnote *
* Values are mean ± standard deviation.
† 1 Japanese yen = 0.007 Euros = 0.008 US dollars in April 2006.
‡ Tests for linear trend used the median value in each quintile as a continuous variable in linear regression.
Table 4 shows the association between monetary cost of dietary energy and dietary intake. Increasing monetary cost of dietary energy was associated with both favourable and unfavourable dietary intake patterns. At the food level, monetary cost of dietary energy was positively associated with consumption of vegetables, fruits, fish and shellfish, pulses, potatoes and dairy products, but also associated positively with intakes of fat and oil, meat and energy-containing beverages, and negatively with intakes of cereals (rice, bread and noodles) (all P for trend <0.01). At the nutrient level, monetary cost of dietary energy was directly associated with intakes of protein, dietary fibre and key vitamins (such as vitamins A, D, E, C, thiamin and riboflavin) and minerals (such as potassium, iron, calcium and magnesium), while monetary cost of dietary energy was associated positively with intakes of fat, saturated fatty acids, cholesterol and sodium, and negatively with carbohydrate intake (all P for trend <0.0001). Monetary cost of dietary energy was positively associated with both energy intake and edible weight consumed (both P for trend <0.0001), but the magnitude of differences between quintiles was larger in edible weight consumed than in energy intake. As a result, monetary cost of dietary energy was negatively associated with energy density (P for trend <0.0001). Adjustment for possible confounding factors did not materially change the associations between monetary cost of dietary energy and dietary intakes (data not shown) except for noodles (P for trend = 0.88), confectioneries (P for trend = 0.0076; negative relationship) and dairy products (P for trend = 0.07).
Table 4 Dietary intake according to quintile category of monetary cost of dietary energyFootnote *
* Values are mean ± standard deviation.
† 1 Japanese yen = 0.007 Euros = 0.008 US dollars in April 2006.
‡ Tests for linear trend used the median value in each quintile as a continuous variable in linear regression. Adjustment for possible confounding variables, including physical activity level (continuous), residential block (Hokkaido and Tohoku; Kanto; Hokuriku and Tokai; Kinki; Chugoku and Shikoku; Kyushu), size of residential area (city with ≥1 million; city with <1 million; town and village), living status (living with family; living alone; living with others), current smoking (yes; no), current alcohol drinking (yes; no), current dietary supplement use (yes; no), currently trying to lose weight (yes; no), and rate of eating (very slow; relatively slow; medium; relatively fast; very fast) and energy intake for variables except for energy intake (kcal day−1, continuous), did not change the results materially with the exception of noodles (P or trend = 0.88), confectioneries (P for trend = 0.0076; negative relationship) and dairy products (P for trend = 0.07).
Table 5 shows the association between monetary cost of dietary energy and BMI. Monetary cost of dietary energy was quite weakly but significantly negatively associated with BMI in both crude (model 1: P for trend = 0.0224) and multivariate (model 2: P for trend = 0.0197) analyses. This negative relationship remained statistically significant after further adjustment for macronutrient and dietary fibre intakes (model 3: P for trend = 0.0301).
Table 5 Body mass index according to quintile category of monetary cost of dietary energy (n = 3931)Footnote *
* Values are mean ± standard error.
† 1 Japanese yen = 0.007 Euros = 0.008 US dollars in April 2006.
‡ Tests for linear trend used the median value in each quintile as a continuous variable in linear regression.
§ Crude model.
¶ Adjusted for physical activity level (continuous), residential block (Hokkaido and Tohoku; Kanto; Hokuriku and Tokai; Kinki; Chugoku and Shikoku; Kyushu), size of residential area (city with ≥1 million; city with <1 million; town and village), living status (living with family; living alone; living with others), current smoking (yes; no), current alcohol drinking (yes; no), current dietary supplement use (yes; no), currently trying to lose weight (yes; no), rate of eating (very slow; relatively slow; medium; relatively fast; very fast) and energy intake (kcal day−1, continuous).
∥ Adjusted for variables used in the model 2 and intakes of protein (% of energy, continuous), fat (% of energy, continuous) and dietary fibre (g 1000 kcal−1, continuous).
Discussion
To date, no information has appeared on the association between dietary costs and diet quality in Japanese populations, among whom rice is a major staple food and whose proportion of energy intake derived from fat is relatively low compared with Western people. In this study of young Japanese women, increasing monetary cost of dietary energy was associated with both favourable and unfavourable dietary intake patterns. Additionally, monetary cost of dietary energy was quite weakly inversely associated with BMI after controlling for possible confounders.
We found that higher monetary cost of dietary energy was associated with higher consumption of not only ‘healthy’ foods such as vegetables, fruits, fish and shellfish, and pulses, but also ‘unhealthy’ foods such as fat and oil, meat and energy-containing beverages. Increasing monetary cost of dietary energy was also associated with decreased intake of cereals, particularly rice. These findings are not consistent with previous Western studies. Diets high in fruits and vegetables and low in fats and sweets were associated with higher diet costs in French adultsReference Drewnowski, Darmon and Briend7. A study of UK women found that high adherence to a healthy dietary pattern was associated with higher monetary costsReference Cade, Upmeier, Calvert and Greenwood6. For Japanese people, cereals (particularly rice) are staple foods and are consumed at almost every meal, accompanied by a main and several side dishes consisting of mainly fish and shellfish, meat, egg, vegetables and pulses. Cereals are relatively inexpensive compared with the component foods of main and side dishes in Japan, as shown in Table 1. It might be suggested that persons with limited money available for foods mainly consume cereals (mainly rice) with a poor amount or variety of main and side dishes, while persons with affordable money for foods increase the amount or variety of main and side dishes with decreasing consumption of cereals (mainly rice). This hypothesis might be supported by the decreasing consumption of rice and increasing consumption of other foods such as meat, vegetables, fish and shellfish, and pulses24 with the observed increase of Gross National Product25 in Japan from 1955 to 2000.
For nutrients, a dietary survey of French adults found that higher diet costs were associated with more nutrient-dense dietsReference Andrieu, Darmon and Drewnowski9. High-fat diets were less expensive than low-fat diets in Danish childrenReference Stender, Skovby, Haraldsdottir, Andresen, Michaelsen and Nielsen5 and French adultsReference Drewnowski, Darmon and Briend7. These findings in Western populations are again inconsistent with our observations. We found that while higher monetary cost of dietary energy was associated with favourable nutrient intake patterns (higher consumption of protein, dietary fibre and key vitamins and minerals), there was also a positive association for dietary fat (and saturated fat and cholesterol) and sodium and a negative association for carbohydrate. The higher fat and lower carbohydrate intake with increasing monetary cost of dietary energy seemed to be due largely to decreasing consumption of cereals (particularly rice). Higher intake of sodium might be due to higher intakes of vegetables, meat, fish and shellfish, and pulses, because in Japan these foods are usually accompanied by seasonings of salty taste, such as salt, soy sauce, miso and dressings; actually, intakes of these foods were positively correlated with sodium intake in the present study (Pearson correlation coefficient = 0.13–0.43; median = 0.33).
Our finding that monetary cost of dietary energy is inversely associated with dietary energy density reflects the results of a number of other studies. Two French studies also found that energy-dense diets are associated with lower diet costsReference Darmon, Briend and Drewnowski8, Reference Andrieu, Darmon and Drewnowski9. Dietary energy density is largely determined by the water content of foodsReference Stubbs, Ferres and Horgan26; unlike packaged energy-dense foods, which are dry and tend to have a stable shelf life (including cereals, confectioneries, and fat and oil)Reference Darmon, Briend and Drewnowski8, transport, storage and wastage costs are all higher for perishable fresh produce (i.e. energy-dilute foods such as fish and shellfish, vegetables and fruit). Thus, the generally observed inverse association between energy cost and energy density can be explained by the fact that, on an energy content basis, energy-dense foods are clearly less costly than energy-dilute foodsReference Darmon, Briend and Drewnowski8. However, although energy density and energy intake have previously constantly been shown to be positively linked, both in experimental and epidemiological studies in Western countriesReference Poppitt and Prentice27, a higher energy intake was associated with a lower energy density because energy intake was positively but energy density was negatively associated with monetary cost of dietary energy in the present study. This seemed to be due to the phenomenon that monetary cost of dietary energy was positively associated with both energy intake and edible weight consumed, and the magnitude of differences between quintiles was larger in edible weight consumed than in energy intake. This important point of disagreement between Western and Japanese studies needs to be addressed in future studies.
A Spanish study reported an association between the higher monetary costs of healthy dietary patterns and lower BMIReference Schroder, Marrugat and Covas10. Although higher monetary cost of dietary energy was not necessarily associated with healthier dietary intake patterns, there was a significant independent negative relationship between monetary cost of dietary energy and BMI in the present study. However, although the association was statistically significant, the magnitude of differences in BMI between quintiles of monetary cost of dietary energy was quite low. Additionally, our subjects were lean compared with Western populations and their BMI values have been reported, not measured, although a high correlation of BMI calculated from self-reported body weight and height with that calculated from measured values has been shownReference Livingstone and Black28, Reference Okubo and Sasaki29, which suggests that BMI calculated from self-reported body weight and height is a reliable measure, at least for use in correlation analysis. Furthermore, although we tried to adjust for a wide range of potential confounding factors, monetary cost of dietary energy might be a surrogate of factors associated with BMI which we did not measure in the present study (e.g. socio-economic level of families), as having or using money is unlikely directly related to BMI. Thus, the public health relevance of this finding is highly uncertain.
A number of methodological limitations of this study should be mentioned. First, in the absence of actual food expenditure data, food prices were derived from the National Retail Price Survey and websites of nationally distributed supermarket and fast-food restaurant chains. As this procedure gives only an approximation of actual diet costs, the results of the present study should be interpreted with caution. We note, however, that a similar methodology has been used in all previous observational studiesReference Stender, Skovby, Haraldsdottir, Andresen, Michaelsen and Nielsen5–Reference Schroder, Marrugat and Covas10.
Second, although we used a validated DHQReference Sasaki, Yanagibori and Amano13–Reference Sasaki, Ushio, Amano, Morihara, Todoriki and Uehara15, the underreporting of dietary intake remains a serious concernReference Lafay, Mennen, Basdevant, Charles, Borys and Eschwege30. To minimise the influence of dietary underreporting, we used energy-adjusted values. Studies have consistently shown that underreporting is more prevalent among people with higher BMIReference Lafay, Mennen, Basdevant, Charles, Borys and Eschwege30, Reference Goodman, Hinden and Khandelwal31, and that energy-dense, nutrient-dilute and low-cost foods such as fat and oils, sugar and confectioneries are more likely to be selectively underreportedReference Lafay, Mennen, Basdevant, Charles, Borys and Eschwege30, Reference Kuczmarski, Kuczmarski and Najjar32. However, as mentioned above, exclusion of energy underreporters identified by the Goldberg cut-off methodReference Goldberg, Black, Jebb, Cole, Murgatroyd and Coward21, Reference Black23 did not change the results materially, which may support the robustness of the findings in the present study.
Several intervention studies in the USA reported nutrient-dense diets that were not more expensive than lower-quality dietsReference Mitchell, Shannon, McKenzie, Smiciklas-Wright, Miller and Tershakovec33–Reference Raynor, Kilanowski, Esterlis and Epstein35. These intervention studies provided individual instruction on how to identify nutritious low-cost foods, how and where to make food purchases, and how to store and prepare the foods, possibly facilitating the consumption of a healthier diet at lower cost. However, the observational nature of the present study did not allow us to investigate directly the cost of dietary change following nutritional intervention.
Finally, because our subjects were selected female dietetic students, we may not be able to extrapolate our results to the general Japanese population. To minimise the influence of nutritional education, the present survey was conducted in most institutions within 2 weeks after the course began. As the subjects are likely to have a high level of nutrition knowledge and elevated rates of social desirability concerning diet, health and body weight, they are also likely to report lower body weights (and higher body heights) and lower energy intakes than average. However, mean values of BMI and energy intake in our subjects (21.0 kg m−2 and 1822 kcal day−1, respectively) were highly comparable with those in a representative sample of Japanese women aged 15–19 years (21.0 kg m−2 and 1858 kcal day−1, respectively)12.
In conclusion, monetary cost of dietary energy was associated with not only favourable but also unfavourable aspects of dietary intake among a large group of young Japanese women. Additionally, monetary cost of dietary energy was inversely associated with BMI, although the magnitude was quite low. Because the relationship between dietary cost, nutrient and food intakes and BMI is an important public health topic, our observation in a selected population should be confirmed using more precise evaluation of diet costs or actual food expenditures in a more representative sample of the Japanese population.
Acknowledgements
Sources of funding: This work was supported by grants from the Japanese Ministry of Health, Labour and Welfare.
Conflict of interest declaration: None of the authors has any conflict of interest to declare.
Authorship responsibilities: K.M. was involved in designing the study, data collection and data management, conducted the statistical analyses, and wrote the manuscript. S.S. was responsible for designing the study, data collection and data management, the overall management, and assisted in manuscript preparation. H.O. was involved in the study design. Y.T. assisted in the manuscript preparation. Y.H. was involved in study design, data collection and data management. M.I. was involved in data collection and data management. All of the authors provided suggestions during the preparation of the manuscript and approved the final version submitted for publication.
