www.elsevier.com/locate/worlddev
World Development Vol. 27, No. 11, pp. 1905±1916, 1999
Ó 1999 Elsevier Science Ltd. All rights reserved
Printed in Great Britain
0305-750X/99/$ - see front matter
PII: S0305-750X(99)00094-7
Urbanization, Lifestyle Changes and the Nutrition
Transition
BARRY M. POPKIN *
University of North Carolina, Chapel Hill, USA
Summary. Ð Broad shifts are occurring at a rapid pace in the structure of diet, physical activity
patterns, and obesity patterns in urban areas in lower income countries. Examples from China and
selected other countries along with pooled time-series and cross-sectional analysis of shifts in diet
and occupation structure associated with urbanization are presented. Patterns of obesity from
nationally representative surveys indicate that the problems of obesity and dietary excess represent
an important challenge facing many lower income countries, particularly their urban populations. Ó 1999 Elsevier Science Ltd. All rights reserved.
Key words Ð nutrition transition, urbanization, obesity, dietary trends
1. INTRODUCTION
The forces that have created major shifts in
fertility and mortality patterns have also been
linked with equally important shifts in diet,
physical activity, and body composition. The
concept of the nutrition transition relates to the
sense that the underlying shifts in economic,
demographic, and related forces that a€ect
fertility, mortality, and disease patterns also
a€ect the structure of diet, physical activity,
and body composition trends. The concept of
the nutrition transition focuses on large shifts
in the structure of diet. The same underlying
socioeconomic and demographic changes
associated with these dietary changes are also
linked with shifts in physical activity, average
stature, and body composition patterns. As we
will show below, urban residency is linked with
large changes in diet and body composition and
also with high levels of obesity in lower and
middle-income countries.
The nutrition transition has followed ®ve
broad patterns: (a) hunting and gathering
food, (b) famine, (c) receding famine, (d)
degenerative disease, and (e) behavioral
change. The major features of each pattern are
summarized in Table 1 (Popkin, 1993). These
patterns are not restricted to particular periods
of human history. For convenience, the
patterns are outlined in past tense as historical
developments; however, ``earlier'' patterns are
not restricted to the periods in which they ®rst
arose, but continue to characterize certain
geographic and socioeconomic subpopulations.
As others have shown (Haddad, Ruel and
Garrett, 1999), many urban populations in the
world still face food insecurity and related
problems. But in the same populations where
undernutrition and food insecurity are found,
other subpopulations, often even in the same
household, su€er dietary excess and obesity. It
is this complex intertwining of economic and
social conditions that creates this rapid transition among some urban subpopulation groups
o€ers a major challenge for food and nutrition
policy in many countries in Asia, Africa, Latin
America, and the Caribbean.
This article summarizes some of the information that documents these broad shifts in the
structure of diet, physical activity patterns, and
obesity patterns in urban areas in lower income
countries. The particular focus is on the
* Preparation
of this article was supported in part by
grants from the US National Institutes of Health (R01HD30880). The author thanks the following sta€ of the
Carolina Population Center, University of North Carolina at Chapel Hill: Tom Swasey for his work on the
graphics, Lynn Igoe for editorial assistance, Frances
Dancy for administrative assistance and Colleen Doak
for her help in preparing the tables. He also credits
collaborators on the China Health and Nutrition
Survey, in particular Drs. Keyou Ge and Fengying
Zhai, Institute of Nutrition and Food Hygiene, Chinese
Academy of Preventive Medicine.
1905
Primitive; onset of ®re
Subsistence; primitive
stone tools
Income and assets
Hunter-gathers
Robust, lean, few
nutritional de®ciencies
Plants, low-fat wild
animals; varied diet
Pattern 1: Collecting
food
Household
production
2. Economy
Nutritional status
1. Nutrition pro®le
Diet
Transition pro®le
Labor-intensive,
primitive technology
begins (clay cooking
vessels)
Subsistence; few tools
Agriculture, animal
husbandry,
home-making begin;
shift to monocultures
Children, women
su€er most from low
fat intake; nutritional
de®ciency diseases
emerge; stature
declines
Cereals predominant;
diet less varied
Pattern 2: Famine
Primitive water
systems; clay stoves;
cooking technology
advance
Increasing income
disparity; agricultural
tools; industrialization
rises
Second agricultural
revolution (crop
rotation, fertilizer);
Industrial Revolution;
women join labor force
Continued MCH
nutrition problems;
many de®ciencies
disappear; weaning
diseases emerge; stature
grows
Fewer starchy staples;
more fruits, vegetables,
animal protein; low
variety continues
Pattern 3: Receding
famine
Rapid growth in
income and income
disparities; technology
proliferation
Fewer jobs with heavy
physical activity;
service sector and
mechanization;
household technology
revolution
Household technology mechanizes and
proliferates
More fat (especially
from animal
products), sugar and
processed foods; less
®ber
Obesity; problems for
elderly (bone health,
etc.); many disabling
conditions
Pattern 4: Degenerative disease
Income growth slows;
home and leisure
technologies increase
Service sector
mechanization,
industrial robotization
dominate; leisure
exercise grows to o€set
sedentary jobs
Food preparation cost
falls signi®cantly with
technological change
Reduced body fat levels
and obesity; improved
bone health
Less fat and
processing; increased
carbohydrates, fruits
and vegetables
Pattern 5: Behavioral
change
Table 1. Summary characteristics of and broad changes in dietary patterns and their relationship to social and economic factorsa
1906
WORLD DEVELOPMENT
a
Rural, low density
Residency Patterns
Reprinted with permission from: Popkin (1993).
Nonexistent
Young population
Age structure
4. Food processing
Much infectious
disease; no epidemics
Low fertility, high
mortality, low life
expectancy
Morbidity
3. Demographic pro®le
Mortality/Fertility
Food storage begins
Rural; a few small,
crowded cities
Epidemics; endemic
disease (plague,
smallpox, polio, TB);
de®ciency disease
begins; starving
common
Young; very few
elderly
Age of Malthus; high
natural fertility, low
life expectancy, high
infant and maternal
mortality
Storage process
(drying, salting);
canning and processing
technologies; increased
food re®ning and
milling
Chie¯y rural; move
to cities increases;
international migration
begins; megacities
develop
Slow mortality
decline, later rapid;
fertility static, then
declines; small,
cumulative population
growth, later explodes
TB, smallpox,
infection, parasitic
disease, polio, weaning
disease (diarrhea,
retarded growth)
expand, later decline
Chie¯y young; shift to
older population begins
Numerous food
transforming
technologies
Life expectancy
hits unique levels
(60s±70s); huge
fertility declines and
¯uctuations (e.g.,
postwar baby boom)
Chronic disease
related to diet,
pollution (heart
disease, cancer);
infectious disease
declines
Rapid fertility decline;
elderly Proportion
increases rapidly
Urban population
disperses; rural green
space reduced
Technologies create
foods and food
constituent substitutes
(e.g., macronutrient
substitutes)
Lower density
cities rejuvenate;
urbanization of rural
areas encircling cities
increases
Increased health
promotion (preventive
and therapeutic); rapid
decline in CHD, slower
change in age-speci®c
cancer pro®le
Increasing proportion
of elderly >75
Life expectancy
extends to 70s, 80s;
disability-free period
increases
URBANIZATION, LIFESTYLE CHANGES AND NUTRITION TRANSITION
1907
1908
WORLD DEVELOPMENT
importance of urban residence in this process.
Unfortunately the analysis is static in that we
do not have many panels that can allow us to
unravel the dynamic changes in urban residence
and its e€ects on diet and activity. We have a
few examples from China. Rather, we must use
crosscountry comparisons along with the
limited longitudinal analysis and our understanding of the urbanization process to infer
anything about the dynamic processes faced by
the world.
2. THE STRUCTURE OF DIET AND
URBANIZATION
People living in urban areas consume diets
distinctly di€erent from those of their rural
counterparts and the general shifts in their diets
enhance energy and fat density and lead to
great potential for chronic disease-related
health problems. A large descriptive literature
on comparisons of urban and rural diets can be
summarized as urban diets show trends toward
consumption of superior grains (e.g., rice or
wheat, rather than corn or millet); more milled
and polished grains (e.g., rice, wheat); food
higher in fat; more animal products; more
sugar; more food prepared away from the
home; and more processed foods (Popkin and
Bisgrove, 1988).
These contrasts between urban and rural
eating patterns are more marked in lower
income than in higher income countries. In
higher income countries, market penetration
into rural areas is common, and national integrated food distribution systems exist.
(a) Crosscountry analysis
Elsewhere we have used data from the Food
and Agriculture Organization (FAO) food
balance sheets for 1962±90, now available in
the FAOSTAT database (Drewnowski and
Popkin, 1997). We combined data on food
availability, expressed in percentage of daily
energy from macronutrients, with the ocial
estimates of gross national product (GNP), as
established by the World Bank (World Bank,
1995). In both cases, GNP per capita was
expressed in 1993 dollars to allow for an easier
interpretation of the results. A basic Ordinary
Least Squares (OLS) regression was used to
relate dietary data (the proportion of energy
from vegetable and animal fats, carbohydrates,
caloric sweeteners, and protein) for those
countries for which full sets of data were
available in 1990 to the logarithm of per capita
GNP. This research used all countries for
which both sets of data were available. These
numbered 133 in 1990. The percentage of
energy from each macronutrient was the
dependent variable and the independent variables were GNP per capita, the proportion of
the population residing in urban areas that
year, and an interaction term between GNP per
capita and the proportion of urban residents.
All variables in this regression were highly
signi®cant …p > 0:01†.
Simulations are used to provide some insight
into the potential e€ect of shifts in the
proportion of the urban population on the
structure of the diet. Urbanization rates were
set at the proportion of urban residents
worldwide at either 25% or 75%. Given rapid
rates of urbanization in developing nations, this
level of urbanization will be seen in most
countries by the year 2020 (United Nations,
1995). As shown in Figures 1 and 2 for higher
rates of urbanization, there is a substantial
increase in the consumption of sweeteners and
fats. The implication is that a shift from 25% to
75% urban population in very low income
countries is associated with an increase of
approximately four percentage points of total
energy from fat and an additional 12 percentage points of energy from sweeteners.
This regression model predicts that rapid
urbanization, usually associated with greater
incomes and economic growth, can have independent e€ects on diet structure. At lower
income levels, according to the regression
model, urbanization can more than double the
amount of sweeteners in the diet. The model
con®rms previous observations that people
living in urban areas consume diets distinct from
those of their rural counterparts (Popkin and
Bisgrove, 1988). The potential impact of
urbanization in ¯attening the income-sweetener
relationship deserves further analysis; however,
it is clear that the increased urbanization of
lower income countries is accelerating the shift
to increased consumption of sweeteners and fats.
Analyzing the impact of urbanization on diet
structure is a key public health issue. Urbanization and economic growth, closely linked
since the industrial revolution, give every sign
of becoming dissociated. In the past rapid
urban growth was concurrent with economic
growth, but now urban growth dominates most
low-income countries.
URBANIZATION, LIFESTYLE CHANGES AND NUTRITION TRANSITION
1909
Figure 1. Relationship between the proportion of energy from each food source and gross national product per capita
with the proportion of the population residing in urban areas placed at 25%, 1990. (Source: Food balance data from the
FAOUN; GNP data from thne World Bank; regression work by UNC-CH Reprinted from Drewnowski and Popkin
(1997) Nutrition Review 55:31.)
Figure 2. Relationship between the proportion of energy from each food source and gross national product per capita
with the proportion of the population residing in urban areas placed at 75%, 1990. (Source: Food balance data from the
FAOUN; GNP data from thne World Bank; regression work by UNC-CH Reprinted from Drewnowski and Popkin
(1997) Nutrition Review 55:31.)
1910
WORLD DEVELOPMENT
Table 2. Percentages of study population aged 20±45 in high and low energy consumed from fat dietary intake categories by tertile of household income, CHNS 1989, 1991, and 1993
Distribution of
sample by %
energy from fat
Household income
Low
1989
Middle
High
1991
1993
1989
1991
1993
1989
1991
1993
% Consuming <10%
Urban
14.3a
Rural
39.2b;c
Total
36.5b; d
10.0b
17.3b
16.4b
1.5a;c
14.7b
12.2b
8.6c
24.7c
18.6c
2.7
9.1
7.1
0.4
9.1
6.5
7.1c
14.8c
11.1c
0.4
3.3
1.8
0.3
3.5
2.3
% Consuming >30%
Urban
19.1
Rural
7.6b;c
Total
8.8b;d
25.4b
12.8b
14.3b
36.4b;c
12.9b
17.3b
19.1d
12.0d
14.6d
45.5
19.9
27.7
51.0
24.9
32.6
22.8c
15.3d
18.9c
62.0
39.1
50.7
66.6
44.1
52.7
a
The proportion di€ers signi®cantly from middle and high-income groups within same year …p < 0:05†.
The proportion di€ers signi®cantly among three income groups within same year …p < 0:05†.
c
The proportion di€ers signi®cantly from corresponding value in other two years …p < 0:05†.
d
The proportion di€ers signi®cantly from corresponding value among the three years …p < 0:05†.
b
(b) China case study
In a large number of descriptive and more
rigorous longitudinal studies, we have found
that diets in urban areas of China were more
rapidly becoming diverse and shifting to what
would be termed the Western-diet, higher in
meats, edible oils, and other fats and re®ned
carbohydrates and lower in ®ber. The China
Health and Nutrition Survey (CHNS), a
longitudinal study of about 4,000 households
from eight provinces of China is used for this
analysis. At each time period three days of
dietary intake were collected from each individual in the family. Combined with weighed
household food intake, this allows us to
develop quite accurate pro®les of each person's
diet (Zhai et al., 1996). More detail on the
surveys is obtained elsewhere (Entwisle et al.,
1995; Popkin et al., 1993).
While the traditional Chinese diet was felt to
be a low-fat one, only a small and rapidly
diminishing proportion of the population now
follows this traditional low-fat pattern and an
ever-increasing proportion is consuming more
than 30% of their energy from fat. This high-fat
diet was signi®cantly more common in urban
and higher income populations than in rural
and lower-income ones (see Table 2). At the
same time, there were decreases in the proportion of adults consuming a low-fat diet among
all income groups.
The e€ects of urban residency on food choices in China have been studied in fully speci®ed
longitudinal models of food choice. These
models of adult food consumption used the ®rst
three waves of the CHNS. When income, food
price, and a range of other sociodemographic
variables are controlled for, we ®nd that there is
still an urban e€ect on the likelihood of
consuming selected food groups as well as on
the consumption of food by those selecting to
consume each food group (Guo et al., 1998)
(see Table 3).
While it is clear from this work that urban
residence is linked with shifts in the structure of
diets toward more higher fat foods, little is
understood abut the dynamics. That is, we do
not have any clear sense if the causes for these
changes are underlying SES factors such as
higher incomes and lower food prices or there is
a broader more synergistic e€ect of all these
factors that creates a unique urban residence
e€ect. In addition, there are no studies on the
rapidity of change to tell us if change is
occurring in di€erent ways in urban and rural
areas.
3. URBANIZATION AND SHIFTS IN
PHYSICAL ACTIVITY PATTERNS
(a) Crossnational trends
A major change in economic structure associated with the nutrition transition is the shift
from a preindustrial agrarian economy to
industrialization. This transformation then
accelerates; the service sector grows rapidly,
industrial production is dominated by capitalintensive processes, and time-allocation
URBANIZATION, LIFESTYLE CHANGES AND NUTRITION TRANSITION
1911
Table 3. Urban coecients
Determinants
a
Rice
Wheat
¯our
Coarse
grains
Red meat
Eggs
Edible oils
A. Results of longitudinal probit analysis for the likelihood of consuming food-groups % in China, 1989±93
0.49
ÿ0:12
0.12
0.00
ÿ0:04
Urban
ÿ0:16
(1,0)
B. Results of longitudinal regression analysis for the quantity of major food-groups consumed (g/d) in China,
1989±93
19.70
ÿ50:45
ÿ2:99
2.28
ÿ1:98
Urban
ÿ65:61
(1,0)
a
By food prices in the state store and annual per capita income.
Indicates the parameter estimates are signi®cantly di€erent from zero at the 1% and 5% levels, respectively.
**
patterns change dramatically. Associated
socioeconomic changes especially important in
the nutrition transition are changes in women's
roles (especially with respect to patterns of time
allocation), in income patterns, in household
food-preparation technology, in food production and processing technology, and in family
and household composition.
The sectoral distribution of the labor force
toward industry and service has accelerated
around the world. Occupational data obtained
from the World Bank are used to study the
structure of the occupations. A priori, one
would expect that there would be few agricultural jobs in urban areas relative to rural
areas. We would also expect a much more
rapid increase in the proportion of service
jobs in urban areas. We ran time-series
regressions for a set of lower-income countries
for 1972±95. The regression equation, using
OLS weighted by the total work force in
each country, produced signi®cant year,
proportion urban, and urban interacted with
year coecients (p > 0:001 for all three sets of
regressions for service, manufacturing, and
agricultural occupations). We simulated the
occupation time trends for two scenariosÐ
25% of the population in the country is urban
and 75% is urban.
The results, as expected, show a much higher
and more rapid shift toward service and away
from agriculture in the urbanized countries (see
Figures 3 and 4). In contrast, the less urbanized
countries are more highly agricultural and have
seen a much slower shift toward service and
manufacturing employment. One of the most
inexorable shifts with modernization and
industrialization is the reduced use of human
energy to produce goods and services. The
result is obviously a marked shift in activity
patterns at work, a trend particularly associ-
ated with our shift into increasingly capital_intensive production and increasingly
sedentary service and commercial work in more
urbanized populations.
(b) China case study
Unfortunately, there are few studies of this
shift in activity and energy expenditures. One
quite simple measure of overall activity has
been collected in each survey of the CHNS. In
Table 4, we grouped physical activity measures
into low, middle, and higher (with low being
the least stressful and active patterns) (Popkin,
1998). Activity patterns for Chinese adults
shifted remarkably during 1989±93. In particular, urban residents in all income groups were
more likely in 1993 to have adopted a more
sedentary activity pattern (Table 4). In
contrast, this pattern was not seen in the rural
areas. In fact, rural residents, particularly lowincome ones, showed a signi®cant change from
low and moderate activity patterns toward a
high physical activity pattern.
4. URBANIZATION AND OBESITY
(a) Crosscountry descriptive work
Body mass index (BMI) is the standard
population-based measure of overweight and
obesity status. This uses height in meters divided by weight in kilograms squared. For adults,
the cuto€s used to delineate obesity are less
than 18.5 for thinness (chronic energy de®ciency), 18.5 to 24.99 for normal, 25.0 to 29.99
for overweight Grade I, 30.0 to 39.99 for
overweight Grade II, and 40.0 and above for
overweight Grade III (WHO, 1995). For this
article, Grades II and III are combined. Ideally
1912
WORLD DEVELOPMENT
Figure 3. Shifts in the distribution of occupations for lower income countries, simulations assuming 25% of the
population living in urban areas, 1965±93. (Source: World Bank, 100 countries ®gure created using simulation of OLS
weighted regression of pooled timed series with urban interaction term.)
Figure 4. Shifts in the distribution of occupations for lower income countries, simulations assuming 75% /of the
population living in urban areas, 1965±93. (Source: World Bank, 100 countries ®gure created using simulation of OLS
weighted regression of pooled timed series with urban interaction term.)
URBANIZATION, LIFESTYLE CHANGES AND NUTRITION TRANSITION
1913
Table 4. Distribution of Physical Activity of Chinese Aged 20±45, by Tertile of Household Income and Residence,
CHNS 1989, 1991, and 1993e
Residence and
activity %
Household income per capita tertiles
Low
Middle
High
1989
1991
1993
1989
1991
1993
1989
1991
1993a
Urban residence
Lowest level
Middle level
Highest level
23.7b; d
49.6b;c
26.7b
34.3b
30.1
35.6b;c
42.6
30.2
27.2
35.5c
46.1c
18.4
45.4
39.7a
14.9
42.8
40.6
16.6
44.5c
48.0c
7.5
58.3
34.5
7.2
55.5a
32.0a
12.5a;c
Rural residence
Lowest level
Middle level
Highest level
15.3c
22.2b;d
62.5b;d
3.9b
5.3b
90.8b
4.8b
7.9b
87.3b
16.2c
28.9c
54.9c
12.3
14.1
73.6
12.6
13.3
74.1
23.7a
35.2c
41.1c
24.8
24.0
51.2
19.8
26.8c
53.4
a
The proportion di€ers signi®cantly among three income groups within same year …p < 0:05†.
The proportion di€ers signi®cantly from corresponding value among three years …p < 0:05†.
d
The proportion di€ers signi®cantly from middle and high-income groups within same year …p < 0:05†.
c
The proportion di€ers signi®cantly from corresponding value in other two years …p < 0:05†.
e
Reprinted with permission from: Popkin (1998).
b
we would follow these cuto€s universally;
unfortunately many published results use
earlier cuto€s (e.g., many use a level for Grade I
of a BMI above 25 and others use the National
Center for Health Statistics percentile cuto€s
that gave males a cuto€ of 27.8 and females
27.3) and the datasets are unavailable for revision.
Nationally representative or large representative surveys from a few countries provide
urban-rural comparisons of the obesity
patterns. More thorough analysis of obesity
patterns and trends is presented elsewhere
(Popkin and Doak, 1998). In general, there are
far higher levels of obesity (especially Grade II
and above) in the Middle East, the Western
Paci®c, and Latin America. In Figure 5 we
provide examples of obesity level di€erences
between urban and rural areas. As expected,
there is a much higher level of obesity in urban
than in rural areas.
(b) China case study
All of the longitudinal analyses of the determinants of obesity in China have used ®xed
e€ects models and have thus removed any e€ect
of residence in an urban area; however, when
urban was interacted with a time-varying
physical activity measure its coecient was
signi®cant (p > 0:001, from unpublished
results). Cross-sectional OLS analysis of the
e€ects of fully speci®ed models that include
urban dummy variables show that urban resi-
dents are far more likely to be obese, even when
controlling for diet and physical activity and
smoking behavior (Paeratakul et al., 1998).
5. HEALTH IMPLICATIONS
A range of changes in health accompany the
nutrition transition in urban areas. Most
immediate seems to be the emerging epidemic
of non-insulin-dependent diabetes mellitus
(NIDDM). There is a growing literature that
documents rapid increases in NIDDM in many
lower-income countries (Hodge et al., 1996,
1997; Levitt et al., 1993; Zimmet, McCarty and
de Courten, 1997). Other work indicates that
many of the other cardiovascular conditions
related to NIDDM, such as hypertension,
dyslipidemia, and atherosclerosis, are documented as increasing rapidly (e.g, Beaglehole,
1992). Recently a most provocative cancer
study has laid a strong basis for linking the diet,
activity, and body composition trends discussed above to the likelihood of increased rates of
prevalence for a larger number of cancers
(World Cancer Research Fund, 1997).
A related clinical and epidemiological literature highlights the importance of the same
factors noted as being central to the nutrition
transitionÐdiet shifts, reduced physical activity, and obesity as also being critical determinants of NIDDM (Zimmet, McCarty and de
Courten, 1997). We review this literature in
more detail elsewhere (Popkin, 1998).
1914
WORLD DEVELOPMENT
Figure 5. Developing country obesity patterns: urban±rural di€erences.
6. DISCUSSION
This article presents a great deal of descriptive data that show that urban areas in lower
and middle-income countries are further along
on the shift toward more ``Western diets''
dominated by more re®ned foods and a higher
fat diet and also a more sedentary lifestyle and
more obesity. The conditions that are leading
to the rapid change in diet, activity, and obesity
are linked with many chronic diseases. Together these e€ects are leading to a rapid shift in
the composition of disease in lower-income
countries. Unlike infectious and childhood
disorders, these chronic diseases are very
medical care-intensive and these emerging
disease patterns have important implications
for health care costs for many decades into the
future. They also can have important e€ects on
productivity, morbidity patterns, and a wide
range of measures of quality of life.
An issue that we have not addressed relates
to the question: is this a rich or poor person's
disease and diet pattern? For lower_income
countries, a crucial dimension of the relationship between socioeconomic status and nutrition is the distribution of chronic disease risk
factors by income group. In particular, a World
Bank study on adult health in Brazil (World
Bank, 1990) indicates that where income
constraints among the poor are not too severe,
many risk factors for cardiovascular disease
will likely be greater among the poor than
among the rich. Monteiro et al. (1998) use 1996
data from Brazil to show that there is an
independent e€ect of economic variables on the
risk of female obesity only in rural areas (the
poorer the woman, the lower the risk of overweight). In the urban context income does not
matter and either formal education or access to
information is each independently and inversely
related to overweight (Monteiro et al., 1998). In
other words, on the national level in urban
Brazil the more educated are less likely to be
overweight than the less educated.
To date, little is understood about the array
of public health and macroeconomic price and
other policy options for addressing these
problems. Moreover as noted earlier, often this
condition of urban dietary excess and obesity
a€ects one segment of the population while
another faces hunger and food insecurity. Our
purpose is not only to understand the dietary
and health changes taking place, but also to
begin to focus attention on de®ning the
program and policy changes that could redirect
the nutrition transition in many regions of the
world.
Much of our focus in food and nutrition and
economic planning produces results that only
enhance this nutrition transition and adverse
set of lifestyle changes. Decades of ®ghting
poverty and of working to increase fat and
animal consumption is one aspect of this
URBANIZATION, LIFESTYLE CHANGES AND NUTRITION TRANSITION
current set of programs and policies that has
led to an entrenched bureaucracy trained to
encourage such activities.
One must go to Scandinavia and Mauritius
to ®nd countries which have attempted to deal
systematically with the transition's negative
e€ects. The example of Norway demonstrates
the diculty of producing major changes in the
structure of the diet (cf. Milio, 1990). Over a
decade of very active food and nutrition policy
implementation in Norway has included
extensive agricultural, health, and nutrition
education, and regulatory sector e€orts to
reduce saturated fat intake. These policies
resulted in a decrease in the proportion of total
and animal fat in the national diet.
In lower income countries, the most positive
case is that of Mauritius. This small island
republic in the Indian Ocean found such a high
level of cardiovascular disease as part of an
adult health survey conducted in 1987 that it
launched a broad comprehensive health
promotion program. The government used the
mass media, price policy, other legislative and
®scal measures, and widespread education
activity in the community, workplace, and
schools. The results were remarkable: hypertension was reduced considerably, mean serum
cholesterol decreased, cigarette smoking by
1915
men and women declined, as did heavy alcohol
use, and there was increased activity (Dowse et
al., 1995; Uusitalo et al., 1996). Obesity levels
continued to increase and there was little
change in the rate of glucose intolerance.
Development of food and nutrition and
health policies for countries where problems of
dietary excess and de®cit exist side by side
represents a new and pressing agenda. This is
particularly the case in the more urban regions
of the world, but as was shown from the limited
data on obesity in a few sub-Saharan Africa
countries, it is even beginning to be true for
that region of the world. In such countries, the
prevailing policies to promote agricultural and
health change to address problems of de®cit are
quite di€erent from those needed to address
problems of excess. In one example of policyrelevant research, Guo et al. (1998) show how
price changes in China could have improved
protein intake and reduced fat intake if careful
selection of prices to be subsidized took place.
The challenge is clear. Surprisingly little is
known and minimal research is focused on this
topic; however, the evidence from large-scale
surveys and other sources indicates that the
lifestyle changes associated with urbanization
are producing major problems linked with
dietary excess.
REFERENCES
Beaglehole, R. (1992) Cardiovascular disease in developing countries: An epidemic that can be prevented.
British Medical Journal 305, 1170±1171.
Dowse, G. K., Gareeboo, H., Alberti, K. G., Zimmet,
P., Tuomilehto, J., Purran, A., Fareed, D., Chitson,
P., and Collins, V. R. (1995) Changes in population
cholesterol concentrations and other cardiovascular
risk fact or levels after ®ve years of the non_communicable disease intervention programme in Mauritius. British Medical Journal 311, 1255±1259.
Drewnowski, A. and Popkin, B. M. (1997) The nutrition
transition: New trends in the global diet. Nutrition
Reviews 55, 31±43.
Entwisle, B., Henderson, G. E., Short, S., Bouma, J. and
Zhai, F. (1995) Gender and family businesses in rural
China. American Sociological Review 60, 36±57.
Guo, X., Popkin, B. M., Mroz, T. A. and Zhai, F. (1998)
Food price policy can signi®cantly reduce the negative e€ects of the nutrition transition in China.
Carolina Population Center, the University of North
Carolina at Chapel Hill, Chapel Hill.
Haddad, L., Ruel, M. and Garrett, J. (1999) Are urban
poverty and undernutrition growing? some newly
assembled evidence. World Development 27(11), this
issue.
Hodge, A. M., Dowse, G. K., Gareeboo, H., Tuomilehto, J., Alberti, K. G. and Zimmet, P. Z. (1996)
Incidence, increasing prevalence, and predictors of
change in obesity and fat distribution over 5 years in
the rapidly developing population of Mauritius.
International Journal of Obesity and Related Metabolic Disorders 20, 137±146.
Hodge, A. M., Dowse, G. K., Toelupe, P., Collins, V. R.
and Zimmet, P. Z. (1997) The association of
modernization with dyslipidaemia and changes in
lipid levels in the Polynesian population of Western
Samoa. International Journal of Epidemiology 26,
297±306.
Levitt, N. S., Katzenellenbogen, J. M., Bradshaw, D.,
Ho€man, M. N. and Bonnici, F. (1993) The prevalence and identi®cation of risk factors for NIDDM
in urban Africans in Cape Town, South Africa.
Diabetes Care 16, 601±607.
Milio, N. (1990) Nutrition Policy for Food±Rich Countries: A Strategic Analysis. The Johns Hopkins
University Press, Baltimore.
Monteiro, C. A., Benicio, M. H., Mondini, L. and
Popkin, B. M. (1998) A shift in the obesity social
paradigm is energing: Economic and cultural-educational predictors of Adult Obesity in Brazil, Mimeo.
1916
WORLD DEVELOPMENT
Paeratakul, S., Popkin, B. M., Ge, K., Adair, L. S. and
Stevens, J. (1998) Changes in diet and physical
activity a€ect the body mass index of Chinese adults.
International Journal of Obesity 22, 424±432.
Popkin, B. M. (1993) Nutritural patterns and transitims.
Population and Development Review 19(1), 138±157.
Popkin, B. M. (1998) The nutrition transition and its
health implications in lower income countries. Public
Health Nutrition 1, 5±21.
Popkin, B. M. and Bisgrove, E. Z. (1988) Urbanization
and nutrition in low-income countries. Food and
Nutrition Bulletin 10 (1), 3±23.
Popkin, B. M. and Doak, C. (1998) The obesity
epidemic is a worldwide phenomenon. Nutrition
Reviews 56, 106±114.
Popkin, B. M., Ge, K., Zhai, F., Guo, X., Ma, H. and
Zohoori, N. (1993) The nutrition transition in China:
a cross±sectional analysis. European Journal of Clinical Nutrition 47, 333±346.
United Nations. Population Division, Department for
Economic and Social Information and Policy Analysis (1995) World Urbanization Prospects: the 1994
Revisions. Estimates and Projections of Urban and
Rural Populations and of Urban Agglomerations. ST/
ESA/SERA/150, United Nations, New York.
Uusitalo, U., Feskens, E. J., Tuomilehto, J., Dowse, G.,
Haw, U., Fareed, D., Hemraj, F., Gareeboo, H.,
Alberti, K. G. and Zimmet, P. (1996) Fall in total
cholesterol concentration over ®ve years in associa-
tion with changes in fatty acid composition of
cooking oil in Mauritius: cross sectional survey.
British Medical Journal 313, 1044±46.
World Bank (1990) Brazil: The New Challenge of
Adult Health. The International Bank for Reconstruction and Development/World Bank, Washington, DC.
World Bank (1995) World Data 1995. CD-ROM. The
International Bank for Reconstruction and Development/World Bank, Washington, DC.
World Cancer Research Fund (in association with
American Institute for Cancer Research) (1997)
Food, Nutrition and The Prevention of Cancer: A
Global Perspective. American Institute for Cancer
Research, Washington, DC.
WHO Expert Committee (1995) Physical Status: The
Use and Interpretation of Anthropometry: Report of a
WHO Expert Committee. WHO Technical Report
Series 854, Geneva: World Health Organization,
Geneva.
Zhai, F., Guo, X., Popkin, B. M., Ma, L., Wang, Q., Yu,
W., Jin, S. and Ge, K. (1996) The evaluation of the
24±hour individual recall method in China. Food and
Nutrition Bulletin 17, 154±61.
Zimmet, P. Z., McCarty, D. J. and de Courten, M. P.
(1997) The global epidemiology of non-insulindependent diabetes mellitus and the metabolic
syndrome. Journal of Diabetes and Its Complications
11, 60±68.