Content uploaded by Lafay Lionel
Author content
All content in this area was uploaded by Lafay Lionel on Jun 24, 2015
Content may be subject to copyright.
The Journal of Nutrition
Nutritional Epidemiology
Nutrient-Dense Food Groups Have High Energy
Costs: An Econometric Approach
to Nutrient Profiling
1,2
Matthieu Maillot,
3–5
Nicole Darmon,
3–5
* Michel Darmon,
3–6
Lionel Lafay
7
and Adam Drewnowski
8
3
INSERM, U476 ‘‘Nutrition Humaine et Lipides’’, Marseille, F-13385 France;
4
INRA, UMR1260, Marseille, F-13385 France;
5
Univ
Me
´
diterrane
´
e Aix-Marseille 2, Faculte
´
de Me
´
decine, IPHM, Marseille, F-13385 France;
6
University of Bordeaux-2, Biochemistry and
Molecular Biology Laboratory, 33076 Bordeaux, F-33076 France;
7
Agence Francxaise de Se
´
curite
´
Sanitaire des Aliments (AFSSA),
Maisons-Alfort, F-94700 France; and
8
Nutritional Sciences Program, School of Public Health and Community Medicine,
University of Washington, Seattle, WA 98195-3410
Abstract
Consumers wishing to replace some of the foods in their diets with more nutrient-dense options need to be able to identify
such foods on the basis of nutrient profiling. The present study used nutrient profiling to rank 7 major food groups and 25
subgroups in terms of their contribution to dietary energy, diet quality, and diet cost for 1332 adult participants in the
French National INCA1 Study. Nutrient profiles were based on the presence of 23 qualifying nutrients, expressed as the
percentage of nutrient adequacy per 8 MJ, and 3 negative or disqualifying nutrients, expressed as the percentage of the
maximal recommended values for saturated fatty acids, added sugar, and sodium per 1.4 kg. Calculated cost of energy (V/8
MJ) was based on the mean retail price of 619 foods in the nutrient composition database. The meat and the fruit and
vegetables food groups had the highest nutritional quality but were associated with highest energy costs. Sweets and
salted snacks had the lowest nutritional quality but were also one of the least expensive sources of dietary energy.
Starches and grains were unique because they were low in disqualifying nutrients yet provided low-cost dietary energy.
Within each major food group, some subgroups had a higher nutritient-to-price ratio than others. However, the fact that
food groups with the more favorable nutrient profiles were also associated with higher energy costs suggests that the
present structure of food prices may be a barrier to the adoption of food-based dietary guidelines, at least by low-income
households. J. Nutr. 137: 1815–1820, 2007.
Introduction
Food prices and diet costs may be one factor limiting the
adoption of healthier diets, especially by the low-income
consumer. That food prices affect food purchases and food
consumption has been repeatedly shown by studies in economics
(1,2), marketing (3), consumer behavior (4,5), and nutritional
epidemiology (6).
If nutrient-poor diets cost less, then economic factors could
help explain the high prevalence of nutrient deficiencies and
nutrition-related diseases, particularly obesity, among the more
disadvantaged populations (7–9). If healthier diets cost more,
then economic barriers may help explain the low consumption
of fruits, vegetables (10,11), and fish (12) among the lower-
income groups. Diet modeling studies using linear programming
suggest that food budget constraints preferentially orient food
choices toward energy-dense diets that are low in essential
nutrients (13,14). In addition, there is accumulating evidence
that the recommended healthier, balanced, or more prudent diets
are associated with higher costs than are the ‘‘unhealthy’’ diets
(15–17). In particular, the consumption of higher amounts of
fruit and vegetables (18) and essential micronutrients (19) has
been associated with higher diet costs, adjusted for energy.
Conversely, high dietary energy density (amount of energy in
100 g of food) has been associated with lower diet costs (20).
A science-based nutrient profiling system, based on the nutri-
tional characteristics of individual foods or food groups, is
currently under consideration by the European Commission
(21). Intended for consumer protection, the system will deter-
mine which foods or categories of foods will be allowed or
disqualified from certain nutritional or health claims (21). Al-
though diverse nutrient profiling schemes are available (22,23),
few have considered the issue of food costs and nutrient-to-price
ratios. The present study tested the hypothesis that the inverse
relationship between nutrient density and energy cost holds not
only between but also within food groups. In both cases, the more
nutrient-dense foods and food categories would be associated
1
Supported by the French National Research Agency’s 2005–2008 Nutritional
Policies project, the French National Institute for Health Prevention and
Education (INPES), and by the USDA Cooperative State Research, Education,
and Extension Service (CSREES), grant 2004-35215-14441.
2
Author disclosures: M. Maillot, N. Darmon, M. Darmon, L. Lafay, and A.
Drewnowski, no conflicts of interest.
* To whom correspondence should be addressed. E-mail: nicole.darmon@
medecine.univ-mrs.fr.
0022-3166/07 $8.00 ª 2007 American Society for Nutrition.
1815
Manuscript received 13 December 2006. Initial review completed 31 January 2007. Revision accepted 11 May 2007.
by guest on June 12, 2013jn.nutrition.orgDownloaded from
with higher costs, whereas the least nutrient-dense foods and
categories would be associated with lower costs. The relationship
between the qualifying (beneficial) and the disqualifying (neg-
ative) nutrients and energy cost was a topic of particular interest
to us. We used an across-the-board nutrient profiling system to
estimate the nutritional quality of food groups, based on 23
qualifying nutrients and 3 disqualifying nutrients, saturated
fatty acids (SFA),
9
added sugars, and sodium. Nutrient profiling
of foods has been recently described as a powerful tool to rank
foodstuffs according to their contribution to a balanced diet (22).
Materials and Methods
Food consumption data. The national INCA dietary survey, con-
ducted in 1999 by the French National Agency for Food Safety, provided
the food consumption data used in this study. This survey was based on a
nationally representative sample of 1985 French adults, aged 15–92 y,
who were selected using the quota method of sampling (24). All
participants completed a 7-d food record, which was aided by a
photographic manual of portion sizes. Subjects who under- or over-
reported their energy intakes (284 men and 312 women) according to the
method of Black (25) were removed from the sample. The physical
activity level assumed in the calculation of the threshold was 1.55,
corresponding to seated work with low walking and leisure activity. The
final sample of 1332 included 596 men (age range of 15 to 92 y) and 736
women (age range of 15 to 90 y).
Drinking water, diet beverages, tea, and coffee were excluded from
all analyses. The nutritional composition of the remaining 619 foods was
computed from the INCA food composition database (26), the Suvimax
food composition database (27), and other databases (28–30) including
the USDA food composition data for zinc, copper, iodine, and selenium
(31). The French mean national 1997 retail prices, mainly obtained from
marketing research (SECODIP), were also added to the analysis. All
prices were adjusted for preparation and waste using conversion factors.
Food groups. The foods were aggregated into 7 major food groups and
25 food subgroups (Table 1) according to the classification system used
to develop the French food-based dietary guidelines (32). The starches
and grains group included grains, starchy vegetables, dry beans, and
peas. The fruit subgroup included fruit juice and other processed fruits;
the vegetables subgroup included frozen and canned vegetables as well as
soups. The sweets subgroup included sweets, chocolate, pastries,
cookies, ice-creams, and desserts; the salted subgroup included chips,
savory snacks, and salted nuts. The mixed dishes subgroup included
foods like couscous-based dishes, paella, and cassoulet (french equiva-
lent of a bean dish with meat); and a snacks subgroup included foods like
pizzas, quiches, and sandwiches.
Contribution of food groups to diet energy and diet cost. Daily
energy intakes (in MJ/d) and daily diet costs (in V/d) were calculated for
each participant. The percentage of contributions of each food group to
total energy intakes and to the estimated diet costs were determined. The
cost of energy was calculated separately for each food group and
subgroup and was expressed in V/8 MJ (i.e., V/1913 kcal). We chose the
8 MJ value because it is close to the recommended energy intakes for the
studied population of French adults: 9.2 MJ (2200 kcal) for inactive men
and 7.5 MJ (1800 kcal) for inactive women.
Nutrient profiling of food groups. Nutrient profiling of the 7 food
groups and 25 subgroups was based on 2 indicators. An expanded and
modified version of a previously used nutrient density score (NDS) (33)
assessed the presence of qualifying nutrients thought to have a beneficial
effect on health. The score, based on 23 nutrients (Table 2), was the
mean of percentages of the French 2001 recommended dietary allow-
ances (RDA) (34) for each nutrient based on 8 MJ (1913 kcal) of the food
group consumed. The NDS algorithm was as follows:
NDS
ik
¼½ð+
P¼23
P¼1
ðNutrient
ikp
=RDA
p
Þ=23Þ 3 100 3 8=EI
ik
where Nutrient
ikp
is the daily content (g, mg, or mg) of nutrient p
provided by group (or subgroup) k to a subject i, and RDA
p
is the French
RDA for nutrient p. EI
ik
is the energy content (in MJ) provided by group
k to a subject i. A NDS of 100% indicates that the consumption of 8 MJ
(i.e., 1913 kcal) of any one food group or subgroup covers a mean of
100% of the RDA for 23 nutrients (34) (Table 2). Only those nutrients
naturally present in foods were included in the calculation of the NDS.
We developed a second indicator of limited nutrients (LIM) specif-
ically for this study. The LIM used 3 negative or disqualifying nutrients,
TABLE 1 Food groups, subgroups and number of foods
per group
Groups n Subgroups
Meat 166 Shellfish, fatty fish, lean fish, poultry, red meat,
organ meats, deli meats, eggs
Fruit and vegetables 114 Vegetables, fruit, dried fruit, nuts
Mixed dishes and snacks 70 Mixed dishes, snacks
Dairy 87 Yogurts, cheese, milk
Starches and grains 37 Legumes, whole grains, refined grains, potatoes
Sweets and salted snacks 107 Sweets, salted snacks
Added fats 39 Animal fats, vegetable fats
9
Abbreviations used: LIM, limited nutrient score; MRV, maximal recommended
value; NDS, nutrient density score; RDA, recommended dietary allowances;
SFA, saturated fatty acids.
TABLE 2 Recommended dietary allowances and maximal
recommended values used to calculate the nutrient
density and the limited nutrient scores
Men Women
Nutrients RDA RDA
Proteins, g/d 70 50
Fiber, g/d 30 30
Linoleic acid, g/d 10 8
Linolenic acid, g/d 21.6
Docosahexaenoic, g/d 0.12 0,10
Vitamin A, mg/d 800 600
Thiamin, mg/d 1.3 1.1
Riboflavin, mg/d 1.6 1.5
Niacin, mg/d 14 11
Vitamin B-6, mg/d 1.8 1.5
Folates, mg/d 330 300
Vitamin B-12, mg/d 2.4 2.4
Ascorbic acid, mg/d 110 110
Vitamin E, mg/d 12 12
Vitamin D, mg/d 55
Calcium, mg/d 900 900
Potassium, mg/d 3100 3100
Iron, mg/d 916
Magnesium, mg/d 420 360
Zinc, mg/d 12 10
Copper, mg/d 2.0 1.5
Iodine, mg/d 150 150
Selenium, mg/d 60 50
Limited nutrients MRV MRV
Saturated fat acids, g/d 25 20
Added sugars, g/d 55 45
Sodium, mg/d 2365 2365
1816 Maillot et al.
by guest on June 12, 2013jn.nutrition.orgDownloaded from
which, when present in a food, could disqualify it from bearing a
nutritional or health claim. Unlike the NDS, the LIM was calculated for
a given quantity, not a given energy content, to avoid favoring energy-
dense foods. We chose a quantity of 1.4 kg, which approximated the
daily intake of foods (excluding alcohol and nonenergetic beverages) in
this population of French adults (1523 g for men and 1302 g for women).
The LIM was calculated as follows:
LIM
ik
¼½ð+
t¼3
t¼1
ðL
ikt
=MRV
t
Þ=3Þ 3 100 3 1400=Q
ik
;
where L
ikt
is the daily amount (in g or mg) of LIM t provided by group
(or subgroup) k to a subject i. MRV
t
is the maximal recommended value
for limited nutrient t (34). Q
ik
is the quantity of foods (in g) from group k
consumed by subject i. The 3 limited nutrients were sodium, simple
added sugars, and SFA. The MRV for SFA and added sugars corre-
sponded to 10% of the recommended energy intake, i.e., 9.2 MJ (2200
kcal) for inactive men and 7.5 MJ (1800 kcal) for inactive women (35).
The MRV for sodium corresponds to a daily intake of 6 g NaCl.
A LIM of 100% would indicate that the consumption of 1.4 kg of
any one food group or subgroup would provide a mean of 100% of the
MRV for sodium, added sugars, and SFA.
Statistical analyses. Differences between means were tested using
ANOVA. Food groups were sorted by decreasing cost of energy, and
decreasing and increasing trends of mean NDS and LIM were respec-
tively tested. All models were adjusted for within-subject effect. Statis-
tical significance was determined at a ¼ 0.05. All analyses were
performed using SAS software, version 9.1 (SAS Institute).
Results
Mean energy intakes were 9.9 MJ/d (2368 kcal/d) for men and
7.8 MJ/d (1866 kcal/d) for women. The daily cost of the
participants’ diets was estimated at 5.26 V/d (i.e., $6.89/d) for
men and 4.26 V/d (i.e., $5.59/d) for women. The mean cost of
the standard daily energy ration of 8 MJ was therefore 4.25 V
(i.e., $2.90/1000 kcal) for men and 4.37 V (i.e., $2.28/1000
kcal) for women.
Contribution of food groups to diet energy and diet cost.
We calculated the contribution of each of the 7 major food
groups to the total energy content and total cost of diets
consumed by the INCA participants as well as the cost of dietary
energy (in V/8 MJ) for each food group (Fig. 1A). Food groups
that contributed the most energy to the population diet were not
those that contributed the most to diet cost. In this population
sample, fruit and vegetables contributed only 8% of the total
dietary energy, but accounted for 17% of total diet cost. The
meat group contributed 18% of total energy intakes but 35% of
diet costs, whereas mixed dishes and snacks contributed 10%
energy and 13% of diet cost. Conversely, starches and grains,
sweets and salty snacks, and added fats contributed much more
dietary energy in relation to cost. Starches and grains accounted
for 23% of dietary energy but only 9% of the daily diet cost,
whereas added fats provided 10% of dietary energy and 2% of
diet cost. Dairy products were well balanced with respect to
dietary energy and diet cost, contributing ;11% to each. This
energy-to-cost relation by food group was illustrated by ranking
the food groups according to their decreasing order of energy
cost (Fig. 1A).
Nutrient profiling and cost of energy of food groups. The
relationship between mean NDS and cost of energy for each of the
7 major food groups is indicated in Figure 1B. The fruit and
vegetables group had the highest mean NDS, covering 210% of
the RDA. For the same quantity of energy (8 MJ), the meat group
covered a mean of 174% of the RDA, whereas the milk group
covered 112%. The remaining food groups all had NDS below
100%, with sweets and salted snacks scoring the lowest (38%).
The relationship between mean limited nutrient score and
cost of energy among the 7 major food groups is shown in Figure
1C. The fruit and vegetables group had the lowest mean LIM,
with 1.4 kg providing only 21% of MRV. For the same quantity,
the starches and grains group covered a mean of 76%. The
remaining food groups all had LIM .100%, with added fats
and sweets and salted snacks scoring the highest (816% and
387%, respectively).
The lower energy cost was paralleled by a lower NDS (P for
trend ¼ 0.01) and by a higher LIM (P for trend ¼ 0.01) (Fig. 1B,
C). These data confirm the hypothesis that higher energy costs
FIGURE 1 Energy and cost contributions (%) to the total energy
content and total cost of diets (A), NDS (B), LIM (C), and dietary energy
cost (A, B, C) of each of the 7 food groups consumed by INCA
participants. P for decreasing and increasing trend of NDS ,0.01 and of
LIM ,0.01, respectively.
Econometric approach to nutrient profiling 1817
by guest on June 12, 2013jn.nutrition.orgDownloaded from
were associated with higher nutritional quality, whereas lower
energy costs were associated with lower nutritional quality. The
mixed dishes, snacks, and dairy products, which had interme-
diate costs of energy, were also intermediate in terms of
nutritional quality: they had relatively high NDS (close to
100%) as well as high LIM (.150%). However, this nutritional
quality-to-price hierarchy among food groups was not absolute.
Starches and grains were unique because they provided dietary
energy at a low cost (2.0 V /8 MJ or $1.37/1000 kcal), without
containing important amounts of disqualifying nutrients. Dairy
products had a better nutritional quality (both a higher NDS and
a lower LIM) than mixed dishes and snacks, although they were
less expensive sources of energy. Sweets and salted snacks were
relatively expensive sources of energy (3.5 V/8 MJ or $2.40/
1000 kcal), given their low NDS (38%) and their high LIM
(387%).
Nutrient profiling and cost of energy for food subgroups.
The NDS and LIM for each of the 25 food subgroups were
ranked according to decreasing cost of energy within each food
group (Table 3). Among the meat group, organ meats had the
highest NDS (754%) and were associated with a low cost of
energy. In contrast, fish and shellfish had NDS that were almost
as high but were also the most expensive in terms of energy cost.
Deli meats had the lowest NDS (120%) and the highest LIM
(454%) in the meat group. Eggs had a good nutritional quality-
to-price ratio insofar as they had the lowest energy cost in this
group for intermediate values of both NDS and LIM.
Vegetables and fruit had NDS .100% and LIM ,100%.
Dried fruits were less expensive as source of energy than fruit
and vegetables, but their nutrient densities were also lower. Nuts
were the least expensive source of energy in this group, but they
also had the highest LIM.
In the dairy group, milk had higher nutritional quality-to-
price ratio than either cheese or yogurt, in that it was the least
expensive source of energy and had both the highest NDS and
the lowest LIM. In the starches and grains group, all subgroups
had low LIM. Legumes also had high NDS (156%). The nutri-
tional quality of sweets and salted snacks was low (low NDS and
high LIM). Within the added fats group, vegetable fats had a
higher nutritional quality-to-price ratio than animal fats: they
had both higher NDS and lower LIM for a lower cost of energy.
Discussion
The data show that food groups and subgroups differ widely in
terms of nutritional quality and in terms of cost per MJ. The
meat and the fruit and vegetables groups that offered the highest
NDS overall were also the most expensive in terms of cost per
MJ. Conversely, added fats provided dietary energy at a very low
cost and had both a low NDS and a high content of negative or
disqualifying nutrients. Mixed dishes, snacks, and dairy pro-
ducts were intermediate in rank, both in terms of nutritional
quality and in terms of cost of energy. Sweets and salted snacks
had a lower nutritional quality than would be expected from
their relatively high cost of energy.
Both fish and vegetables and fruit had good nutrient profiles,
as indicated by very high NDS and by low LIM. However, they
were also associated with higher costs per MJ and therefore with
higher diet costs. On the other hand, as our previous studies
showed (33), vegetables and fruit provided an affordable package
of nutrients (as opposed to energy) per unit cost.
Overall, starches and grains had very favorable nutritional
quality-to-price ratio. These foods appear to be a good choice,
particularly whole or unrefined staples, which provide adequate
nutrition at a moderate cost. Whole-grain cereals generally provided
twice the amount of nutrients than refined cereal products, but at
twice the price. It will be interesting to determine whether the food
choices made by lower income and food insecure persons, high in
grains and starches and low in vegetables and fruit (7), is a rational
behavior in response to economic constraints, or whether tradition
and education are mainly involved in these choices.
Although a clear ranking of nutrient-to-price ratios was
found among food groups, food subgroups showed more
diversity. Although several food subgroups had a high nutri-
tional quality, they were not the most expensive ones within
their group. These subgroups, particularly milk, organ meats,
and eggs, had a very good nutritional quality-to-price ratio.
Vegetable fats, dried fruit, and nuts also showed good nutri-
tional quality-to-price ratios. Interestingly, diets obtained using
a computer to attain the whole set of nutritional recommenda-
tions at the lowest cost preferentially contained foods belonging
to the groups and subgroups identified in the present study as
having good nutritional quality-to-price ratios (36). This does
TABLE 3 Nutrient density score, limited nutrient score, and
cost of energy of food subgroups consumed by
INCA participants
1
Groups and
subgroups n
2
NDS %/8 MJ LIM %/1.4 kg
Cost of energy
V/8 MJ
Meat
Shellfish 370 643 6 373 64 6 28 33.2 6 20.4
Lean fish 921 375 6 92 60 6 32 15.8 6 8.8
Fatty fish 454 622 6 188 130 6 100 14.4 6 8.1
Red meat 1297 147 6 30 138 6 42 10.3 6 3.1
Poultry 1051 168 6 62 63 6 30 8.8 6 2.6
Organ meats 262 754 6 551 83 6 42 7.4 6 3.4
Deli meats 1102 120 6 75 454 6 103 4.8 6 2.0
Eggs 889 212 6 20 139 6 32 3.2 6 0.3
Fruit and vegetables
Vegetables 1318 352 6 172 27 6 14 15.1 6 8.1
Fruit 1220 134 6 50 13 6 22 7.7 6 3.2
Dried fruit 63 85 6 32 4.7 6 1.9 3.9 6 0.9
Nuts 173 120 6 23 145 6 40 1.7 6 0.9
Mixed dishes and snacks
Mixed dishes 967 106 6 40 155 6 50 6.7 6 4.6
Snacks 1027 80 6 18
229 6 77 5.4 6 2.6
Dairy
Yogurts 1040 119 6 31 65 6 43 4.7 6 1.0
Cheese 1238 101 6 15 478 6 66 4.7 6 1.4
Milk 913 138 6 8.7 29 6 5.9 2.1 6 0.2
Starches and grains
Legumes 405 156 6 24 24 6 23 3.6 6 0.9
Whole grains 146 83 6 10 98 6 24 3.5 6 0.9
Potatoes 1243 75 6 12 58 6 38 1.6 6 0.6
Refined grains 1326 40 6 2.8 77 6 21 1.5 6 0.3
Sweets and salted snacks
Salted snacks 493 80 6 99 422 6 150 6.2 6 14
Sweets 1326 37 6 15 388 6 167 2.8 6 1.1
Added fats
Animal fats 1330 25 6 2.9 1040 6 185 1.2 6 0.1
Vegetable fats 1325 80 6 13 360 6 52 0.6 6 0.1
1
Values are means 6 SD.
2
For each subgroup, n indicates the number of participants consuming foods from this
subgroup.
1818 Maillot et al.
by guest on June 12, 2013jn.nutrition.orgDownloaded from
not mean that low-income consumers should select only grains
and starches and stay away from fruit, vegetables, and fish. On
the contrary, the good quality-to-price ratio of grains and
starches leaves a substantial amount of the budget for high-cost,
nutrient-dense foods such as fruit, vegetables, and fish. Model-
ing studies using both cost and nutritional constraints showed
that including important amounts of unrefined starches in the
diet actually made it possible to fulfill all nutritional require-
ments for people on a moderate food budget (36). Interestingly,
such modeled diets also included important amounts of fruit,
vegetables, and fish.
The analysis of the link between diet cost and nutritional
quality has been hampered for a long time by methodological
limitations. Economists, who typically analyze household bud-
gets surveys, lack information about individual consumption
and about the nutritional composition of purchased foods.
Conversely, nutritional epidemiologists lack information on the
price of foods actually consumed by individuals. Associating a
mean price to foods in food consumption surveys (as well as the
mean nutritional composition associated with them) has allowed
investigators to solve this methodological issue and to estimate
the daily cost of each individual diet (15). Although this
approach only roughly estimates individual expenditures, it
seemed valid, in our study, to evaluate mean expenditures for
food consumed at home, insofar as the mean daily cost estimated
from the present data (4.7 V/d or $6.20/d) was very close to that
from the last French household budget survey (37). The price of
a given food varies according to stores, season, brand, size,
packaging, and according to whether it is prepared at home or
bought ready to eat. The use of a mean price partially hid this
variability; although frequently consumed foods weighted
higher in the mean price calculation. For instance, the mean
price of green beans was closer to the price of the processed items
rather than to that of the fresh ones. Likewise, the mean price of
a given fruit was closer to the price in full season rather than to
the price out of season.
Price variability within a single category of food may alter the
nutritional quality-to-price ratio of foods considered individu-
ally. Actually, a British study showed that branded foods
generally cost 2.5 times the price of economy-line foods, but
do not contain more nutrients, so that the quantity of nutrients
bought for one shilling of food was always clearly higher with
economy-line products (38). We considered that this intrafood
cost variability would not alter the nutritional quality-to-price
hierarchy among main food groups, but this requires further
investigation. Another possible drawback was the evolving
nature of the indicators used to estimate the nutritional quality
of food groups. The present NDS was based on 23 nutrients with
a known RDA. Although only some of these nutrients are
implicated in public health problems, the European Commission
takes into account those nutrients that are scientifically recog-
nized as having an effect on health. That list is still not finalized,
especially insofar as nutritional problems are not the same in all
countries because of different food habits, availability, and
different enrichments and supplementation practices. We there-
fore preferred a more universal score than a country-specific
score. On the other hand, one could also argue that our score
does not consider enough different nutrients. Actually, several
bioactive compounds, including polyphenols and some trace
elements, were not included in the NDS, either because the
nutrient composition database was not available or because the
nutritional requirement was not yet defined. Furthermore, we
calculated only those nutrients naturally present in foods and
not those introduced by enrichment. This was done to avoid
direct comparisons between a fortified food and a nonfortified
food with a similar nutrient content.
The nutritional quality-to-price hierarchy presently found
between food groups probably explains the positive association
observed between the nutritional quality of the diet and its cost
(15,18–20). Notwithstanding, the wide disparity of nutritional
quality and prices observed within food groups is compatible
with the fact that improving diet quality is not necessarily
associated with increased diet costs in intervention studies
implicating nutrition education (39–41). Our results suggest
that, by preferentially selecting subgroups that have the highest
nutritional quality-to-price ratio, healthy diets can be obtained
at a moderate cost. However, such low-cost nutritionally
adequate diets (39–41) deviated considerably from the typical
food habits of the population (36). Although nutrition education
could make such diets more attractive, they may not be palatable
enough or socially acceptable. In addition, there is a threshold
cost under which it is impossible to obtain a nutritionally
adequate diet, estimated at ;3.5 V/d per adult in France (36)
and at $116/wk for a 4-person family in the U.S. (42). Many
studies have emphasized that food budgets of the poor are often
under this threshold (16,17,36,43). The fact that food groups
with the more favorable nutrient profile were also the more
expensive sources of energy suggests that the present structure of
food prices does not favor the adoption of food-based dietary
guidelines, at least by low-income people.
Although nutritionally balanced diets can be obtained at
limited cost (36,39–41), often they are neither palatable nor
convenient. It is a major challenge for public health nutrition to
link public health imperatives with economic realities of life in
ensuring that nutritionally adequate and socially acceptable
foods are affordable and available to all. A refinement of food
and agriculture policies and food assistance programs is one
potential strategy for change (44–46, and unpublished data by
Z. Rambeloson, N. Darmon, and E. L. Ferguson). Effective
dietary guidance must take into account both the nutrient profile
of foods and their nutrient and energy costs. These consider-
ations will allow consumers to identify and select optimal diets
at an affordable cost.
Literature Cited
1. Basiotis P, Brown M, Johnson SR, Morgan KJ. Nutrient availability,
food costs, and food stamps. Am J Agric Econ. 1983;65;686–93.
2. Putnam JJ, Allshouse J, Kantor LS. US per capita food supply trends:
more calories, refined carbohydrates, and fats. Food Review. 2002;25:
2–15.
3. Lennernas M, Fjellstrom C, Becker W, Giachetti I, Schmitt A, Remaut
de Winter AM, Kearney M. Influences on food choice perceived to be
important by nationally representative samples of adults in the
European Union. Eur J Clin Nutr. 1997;51:S8–S15.
4. Cabanac M. Palatability vs. money: experimental study of a conflict of
motivations. Appetite. 1995;25:43–9.
5. French SA. Pricing effects on food choices. J Nutr. 2003;133:841S–3S.
6. Drewnowski A, Darmon N. Food choices and diet costs: an economic
analysis. J Nutr. 2005;135:900–4.
7. Caillavet F, Darmon N, Lhuissier A, Re
´
gnier F. L#alimentation des
populations de
´
favorise
´
es en France: synthe
`
se des travaux dans les
domaines e
´
conomique, sociologique et nutritionnel. Les travaux 2005–
2006 de l#Observatoire National de la Pauvrete
´
et de l# Exclusion
Sociale, Paris: La documentation Francxaise; 2006 [cited 2007 May 19].
Available from: http://lesrapports.ladocumentationfrancaise.fr/BRP/
064000163/0000.pdf.
8. James WP, Nelson M, Ralph A, Leather S. Socioeconomic determinants
of health. The contribution of nutrition to inequalities in health. BMJ.
1997;314:1545–9.
Econometric approach to nutrient profiling 1819
by guest on June 12, 2013jn.nutrition.orgDownloaded from
9. Drewnowski A, Darmon N. The economics of obesity: dietary energy
density and energy cost. Am J Clin Nutr. 2005;82:265S–73S.
10. Dibsdall LA, Lambert N, Bobbin RF, Frewer LJ. Low-income consumers#
attitudes and behaviour towards access, availability and motivation to
eat fruit and vegetables. Public Health Nutr. 2003;6:159–68.
11. Giskes K, Turrell G, Patterson C, Newman B. Socio-economic differ-
ences in fruit and vegetable consumption among Australian adolescents
and adults. Public Health Nutr. 2002;5:663–9.
12. Trondsen T, Scholderer J, Lund E, Eggen AE. Perceived barriers to con-
sumption of fish among Norwegian women. Appetite. 2003;41:301–14.
13. Darmon N, Ferguson EL, Briend A. A cost constraint alone has adverse
effects on food selection and nutrient density: an analysis of human diets
by linear programming. J Nutr. 2002;132:3764–71.
14. Darmon N, Ferguson E, Briend A. Do economic constraints encourage
the selection of energy dense diets? Appetite. 2003;41:315–22.
15. Cade J, Upmeier H, Calvert C, Greenwood D. Costs of a healthy diet:
analysis from the UK Women’s Cohort Study. Public Health Nutr. 1999;
2:505–12.
16. Nelson M, Dick K, Holmes B. Food budget standards and dietary
adequacy in low-income families. Proc Nutr Soc. 2002;61:569–77.
17. Vozoris N, Davis B, Tarasuk V. The affordability of a nutritious diet for
households on welfare in Toronto. Can J Public Health. 2002;93:36–40.
18. Drewnowski A, Darmon N, Briend A. Replacing fats and sweets with
vegetables and fruit – a question of cost. Am J Public Health. 2004;
94:1555–9.
19. Andrieu E, Darmon N, Drewnowski A. Low-cost diets: more energy,
fewer nutrients. Eur J Clin Nutr. 2006;60:434–6.
20. Darmon N, Briend A, Drewnowski A. Energy-dense diets are associated
with lower diet costs: A community study of French adults. Public
Health Nutr. 2004;7:21–7.
21. The European Parliament and the Council of the European Union.
Regulation (EC) No 1924/2006 of the European Parliament and of the
Council of 20 December 2006 on nutrition and health claims made on
foods. Official Journal of the European Union. 2006;L404:9–25.
22. Azais-Braesco V, Goffi C, Labouze E. Nutrient profiling: comparison and
critical analysis of existing systems. Public Health Nutr. 2006;9:613–22.
23. Drewnowski A. Concept of a nutritious food: toward a nutrient density
score. Am J Clin Nutr. 2005;82:721–32.
24. Volatier J-L. Enque
ˆ
te INCA (individuelle et nationale sur les con-
sommations alimentaires). AFSSA, Agence Francxaise de Se
´
curite
´
Sanitaire des Aliments Ed. Paris: Lavoisier, 2000.
25. Black AE. Critical evaluation of energy intake using the Goldberg cut-
off for energy intake:basal metabolic rate. A practical guide to its
calculation, use and limitations. Int J Obes Relat Metab Disord. 2000;
24:1119–30.
26. Favier J, Ireland-Ripert J, Toque C, Feinberg M. CIQUAL. Re
´
pertoire
ge
´
ne
´
ral des aliments. Table de composition. Paris: Lavoisier, Tec&Doc,
1995.
27. Ouvrage collectif. Table de composition des aliments SU.VI.MAX.
Paris: INSERM/Economica; 2006.
28. Lamand M, Tressol J, Ireland-Ripert J, Favier J, Feinberg M. CIQUAL.
Re
´
pertoire ge
´
ne
´
ral des aliments. Vol. 4. Table de composition mine
´
rale.
Paris: Lavoisier, Tec&Doc, 1996.
29. Souci SW, Fachmann W, Kraut H. Food composition and nutrition
tables. 6th revised ed. Stuttgart: Medpharm, Scientific Publishers, CRC
Press, 2000.
30. Food Standard Agency. McCance and Widdowson’s. The composition
of foods. 6th ed. Cambridge: Royal Society of Chemistry. 2002.
31. USDA, Agricultural Research Service. USDA National Nutrient Data-
base for Standard Reference. 2006 [cited 2007 May 19]. Available from:
http://www.nal.usda.gov/fnic/foodcomp/search/.
32. Programme National Nutrition Sante
´
. La sante
´
vient en mangeant.
Le guide alimentaire pour tous. INPES. 2002 [cited 2007 May 19].
Institut National de Pre
´
vention et d’Education a
`
la Sante
´
. Available
from: www.lasantevientenmangeant.inpes.sante.fr.
33. Darmon N, Darmon M, Maillot M, Drewnowski A. A nutrient density
standard for vegetables and fruits: nutrients per calorie and nutrients
per unit cost. J Am Diet Assoc. 2005;105:1881–7.
34. Martin A., ed. Apports nutritionnels conseille
´
s pour la population
francxaise. Paris: Lavoisier, 2001.
35. Euro Diet Working Group. Nutrition and diet for healthy lifestyles in
Europe, science and policy implications, 1998–2000. Core report. 2000
[cited 2007 May 19]. Available from: http://ec.europa.eu/health/ph_
determinants/life_style/nutrition/report01_en.pdf.
36. Darmon N, Ferguson EL, Briend A. Impact of a cost constraint on
nutritionally adequate food choices for French women: an analysis by
line
ar programming. J Nutr Educ Behav. 2006;38:82–90.
37. Andrieu E, Caillavet F, Momic M, Lhuissier A, Regnier F. L#alimen-
tation comme dimension spe
´
cifique de la pauvrete
´
. Approches croise
´
es
de la consommation alimentaire des populations de
´
favorise
´
es. Les travaux
2005–2006 de l#Observatoire National de la Pauvrete
´
et de l#Exclusion
Sociale, Paris: La documentation Francxaise. 2006 [cited 2007 May
19]. Available from: http://lesrapports.ladocumentationfrancaise.fr/BRP/
064000163/0000.pdf.
38. Cooper S, Nelson M. ‘‘Economy’’ line foods from four supermarkets
and brand name equivalents: a comparison of their nutrient contents
and costs. J Hum Nutr Diet. 2003;16:339–47.
39. Raynor HA, Kilanowski CK, Esterlis I, Epstein LH. A cost-analysis of
adopting a healthful diet in a family based obesity treatment program.
J Am Diet Assoc. 2002;102:645–56.
40. Burney J, Haughton B. EFNEP: a nutrition education program that
demonstrates cost-benefit. J Am Diet Assoc. 2002;102:39–45.
41. Mitchell DC, Shannon BM, McKenzie J, Smiciklas-Wright H, Miller
BM, Tershakovec AM. Lower fat diets for children did not increase food
costs. J Nutr Educ. 2000;32:100–3.
42. Lino M. Cost changes in the thrifty Food Plan: January 2004 to January
2005. Nutrition Insight. 2005 June;31:2p. Available from: http://www.
cnpp.usda.gov/Publications/NutritionInsights/Insight31.pdf.
43. Morris JN, Donkin AJ, Wonderling D, Wilkinson P, Dowler EA. A
minimum income for healthy living. J Epidemiol Community Health.
2000;54:885–9.
44. Herman DR, Harrison GG, Jenks E. Choices made by low-income
women provided with an economic supplement for fresh fruit and
vegetable purchase. J Am Diet Assoc. 2006;106:740–4.
45. Taylor J. Updating the WIC food packages: it’s about time. National
Health Policy Forum, Issue Brief. 2006 [cited 2007 May 19];816:1–14.
Available from: http://www.nhpf.org/pdfs_ib/IB816_WICFoodPackage_
11-02-06.pdf.
46. Akobundu UO, Cohen NL, Laus MJ, Schulte MJ, Soussloff MN.
Vitamins A and C, calcium, fruit, and dairy products are limited in food
pantries. J Am Diet Assoc. 2004;104:811–3.
47. Rambeloson Z, Darmon N, Ferguson EL. Linear programming can help
identifying practical solutions to improve the nutritional quality of food
aid in France. Public Health Nutr. In press 2007.
1820 Maillot et al.
by guest on June 12, 2013jn.nutrition.orgDownloaded from
