Context Malnutrition causes death and impaired health in millions of children.
Existing interventions are effective under controlled conditions; however,
little information is available on their effectiveness in large-scale programs.
Objective To document the short-term nutritional impact of a large-scale, incentive-based
development program in Mexico (Progresa), which included a nutritional component.
Design, Setting, and Participants A randomized effectiveness study of 347 communities randomly assigned
to immediate incorporation to the program in 1998 (intervention group; n =
205) or to incorporation in 1999 (crossover intervention group; n = 142).
A random sample of children in those communities was surveyed at baseline
and at 1 and 2 years afterward. Participants were from low-income households
in poor rural communities in 6 central Mexican states. Children (N = 650)
12 months of age or younger (n = 373 intervention group; n = 277 crossover
intervention group) were included in the analyses.
Intervention Children and pregnant and lactating women in participating households
received fortified nutrition supplements, and the families received nutrition
education, health care, and cash transfers.
Main Outcome Measures Two-year height increments and anemia rates as measured by blood hemoglobin
levels in participating children.
Results Progresa was associated with better growth in height among the poorest
and younger infants. Age- and length-adjusted height was greater by 1.1 cm
(26.4 cm in the intervention group vs 25.3 cm in the crossover intervention
group) among infants younger than 6 months at baseline and who lived in the
poorest households. After 1 year, mean hemoglobin values were higher in the
intervention group (11.12 g/dL; 95% confidence interval [CI], 10.9-11.3 g/dL)
than in the crossover intervention group (10.75 g/dL; 95% CI, 10.5-11.0 g/dL)
who had not yet received the benefits of the intervention (P = .01). There were no differences in hemoglobin levels between the
2 groups at year 2 after both groups were receiving the intervention. The
age-adjusted rate of anemia (hemoglobin level <11 g/dL) in 1999 was higher
in the crossover intervention group than in the intervention group (54.9%
vs 44.3%; P = .03), whereas in 2000 the difference
was not significant (23.0% vs 25.8%, respectively; P =
.40).
Conclusion Progresa, a large-scale, incentive-based development program with a
nutritional intervention, is associated with better growth and lower rates
of anemia in low-income, rural infants and children in Mexico.
More than half of the yearly 10.8 million deaths of children younger
than 5 years are attributed to malnutrition,1 as
assessed by underweight (≥2 SDs below the weight expected for that age,
according to the international reference recommended by the World Health Organization
[WHO]2). These deaths are not caused by higher
frequency of common childhood diseases but by higher case fatality rates3,4 and would not occur if the children
were not malnourished. Malnourished children who survive have a high risk
of impaired health and function throughout life, which contributes to the
intergenerational continuation of poverty.5 In
developing countries more than one quarter of all children younger than 5
years, about 150 million total, are estimated to be malnourished.6 Available nutritional interventions and technologies
have proven, under controlled conditions, to be efficacious in preventing
and controlling malnutrition7-10 and
improving survival.11 Because large-scale programs
do not provide similar controlled conditions, we would expect these strategies
to have less impact when implemented in such programs.12,13 Inadequate
evaluations to date, however, have meant that little information is available
on the effectiveness of these strategies implemented through large-scale programs.14,15 Reliable evidence of the effectiveness
of these strategies applied in large-scale programs is needed for planning
sound investments in nutrition and shaping health policies that aim to improve
the nutrition and survival of children in countries where the prevalence of
malnutrition remains high.
Despite Mexico's 40-year tradition of high expenditure in food distribution
programs,16 according to a national probabilistic
survey conducted in 1998, there were still high prevalences of stunting (17.7%)
and anemia (27.2%) in children younger than 5 years.17In
this survey, rural children at the lower socioeconomic tertile (the group
equivalent to the beneficiaries of the program discussed herein) had a much
higher prevalence of stunting (38.8%) and a slightly higher prevalence of
anemia (29.9%) than the national averages or the prevalences for all rural
areas (31.6% for stunting and 29.5% for anemia) (J.A.R., unpublished data).
Moreover, the decline observed in stunting between 1988 and 1998 in
Mexico is unsatisfactory relative to advances made in other Latin American
countries of comparable economic conditions.17 Little
can be postulated about the reason for this discrepancy between investments
made and the nutritional outcomes in Mexico because, before 1998, nutrition
interventions were not evaluated. In 1997, a large-scale human capital development
program called "Progresa" (Program for Education, Health, and Nutrition),
which included a nutritional component, was introduced. Evaluation of overall
impact was incorporated into the design of the program after its inception
in 1997 and includes several expected outcomes in health, education, and nutrition.
This article documents the short-term impact of Progresa on nutritional outcomes.
Progresa (currently called Oportunidades) is a large-scale incentive-based
welfare program that has been applied by the Mexican federal government since
1997. Its principal aim is to develop human capital in low-income households.
In other words, the program's underlying premise is that investments in nutrition,
health, and education in young children who live in low-income environments
will improve their chances of accessing higher education, better jobs, and
in turn a better quality of life in adulthood.18 Investment
in these 3 basic needs is considered central to breaking the intergenerational
continuation of poverty.
The incorporation of families across the country into the program had
to be phased in according to the federal resources made available, which in
turn allowed for the ethical quantitative comparison of control and intervention
groups. During its first 3 years, the program succeeded in covering almost
all eligible families living in rural areas. Coverage expanded from about
300 000 families in 1997 to approximately 2.6 million families in 2000,
about 40% of rural families and 10% of all families in Mexico. Currently,
the program covers 4.5 million low-income families (about 20% of all families
in Mexico) in urban and rural areas.
The description of eligibility and program administration that follows
refers to the program in rural areas during the same period when the evaluation
reported in this article was conducted (1998-2000).
Household eligibility for Progresa was determined in 2 stages, first
by identifying underprivileged communities and then by choosing low-income
households within those communities.19 Underprivileged
communities were identified according to the proportion of households living
in poverty by using data from the 1995 National Census. Low-income households
were selected according to an index of objective, easy-to-collect characteristics
such as housing materials, water and sanitation facilities, education, and
family structure; these were shown to be good proxies for poverty assessment.20 To improve transparency, the list of beneficiaries
was approved by a community assembly. On average, 78% of the households in
selected communities were classified as eligible for program benefits, and
97% of these eligible households enrolled in the program.
Progresa provided micronutrient-fortified foods for women and children
and health services and cash transfers for the family. The food supplements
were targeted to the groups of individuals that are more likely to benefit
from the product21: children aged 4 to 23 months,
underweight children aged 2 to 4 years, and pregnant and lactating women in
the low-income beneficiary households. The supplements were designed by a
group of experts in nutrition according to evidence about the nutritional
status of the population for whom the supplements were intended.22,23 The
main ingredients were dry whole milk, sugar, maltodextrin, and the micronutrients
most deficient in their diet (Box).24 Supplements came in the form of
a dry mix offered in 3 flavors that required hydration before consumption.
They were distributed at health centers in 240-g packages containing about
5 daily rations (the daily ration recommended was 44 g). Mothers were instructed
to add 4 spoons of boiled water to 1 ration, which produces a puree consistency,
referred to in Spanish as papilla.
Protein, 5.8 g
Energy, 194 kcal
Fats, 6.6 g
Carbohydrates,
27.9 g
Sodium, 24.5 mg
Iron, 10 mg
Zinc, 10 mg
Vitamin
A, 400 µg
Vitamin E, 6 mg
Vitamin C, 40 mg
Vitamin
B12, 0.7 µg
Folic acid, 50 µg
Families enrolled in the program (Progresa families) received 2 types
of cash transfers every 2 months: a universal cash amount for all families
and a specific cash transfer associated with school attendance of their school-aged
children enrolled in third-grade primary school to third-grade secondary school.
The transfer associated with school attendance varied according to the number
of children attending school and their grade level; in secondary schoolchildren,
it was more for girls than for boys. Families received the universal cash
transfer as long as they complied with specific health care appointments in
health centers for all family members, including immunizations, well baby
care and growth monitoring of children, prenatal and postnatal care and education
for women, check-up visits for other family members, and a mandatory session
on nutrition and health education. Only about 1% of households were denied
the cash transfers for noncompliance during the evaluation period. Monthly
transfers averaged about $25 per family. Typically, the cash transfers added
about 20% to 30% to the household income.
Evaluation Design and Sample
The evaluation of the nutritional impact of the program was commissioned
by the Ministry of Health to the National Institute of Public Health. Planned
as a randomized intervention,25,26 the
study was conducted in 6 contiguous states in the central region of the country
(Guerrero, Hidalgo, Puebla, Querétaro, San Luis Potosí, and
Veracruz), representing the largest area in which the program operated.
Figure 1 shows the sampling
frame and sample sizes. Five hundred six communities were randomly selected
from the pool of communities that were eligible to receive the program benefits;
320 of these communities were randomly assigned to receive Progresa in 1998,
whereas the rest of the communities (n = 186) were assigned to act as controls
for 2 years, that is, their enrollment was scheduled to begin in 2001 to allow
for 2 years without the intervention. However, the enrollments of these 186
communities in the program occurred 1 year earlier than originally planned,
in late 1999, because of political pressure to accelerate the program.
The nutritional impact substudy was conducted in a random selection
of 205 of the 320 communities scheduled to enroll in the program at the end
of 1998 and 142 communities randomly selected from the 186 communities that
enrolled a year later, in late 1999. The 205 communities are referred to as
intervention communities for their earlier and longer exposure time, and the
142 communities that were enrolled after 1 year of observation are referred
to as crossover intervention communities.
A subsample was chosen because a smaller number of children was adequate
to assess the expected 2-year impact on growth (0.8 cm) and anemia (10-percentage-point
reduction). The sample size was calculated according to the original 2-year
intervention design for 1-tailed tests, assuming a .05 significance level
and a power of 90%. It was first estimated as a simple random sample, which
was further multiplied by a design effect of 1.4 to take into account the
complex sample design.27,28
Data on a cohort of all infants 12 months or younger (n = 795; 461 in
intervention and 334 in crossover intervention communities) were collected
at baseline, between August and September 1998, from a random sample of eligible
households with children younger than 5 years. They were surveyed again in
September to December 1999 and in November to December 2000.
To assess differences in height, children were followed up for approximately
2 years, as was planned when the evaluation was designed, from 1998 to 2000.
During this period, the intervention children received 2 years of Progresa,
whereas the crossover intervention children received it during the second
year only. Because all the children were the same age, the intervention children
received Progresa for the first and second year of life, whereas the crossover
intervention children received it in the second year of life only. Thus, we
expected better height growth in the intervention children than in the crossover
intervention children because of a difference in span of exposure to Progresa
and because the crossover intervention children received Progresa only at
an age when the impact of supplementary feeding in height is smaller.21 Because we expected that the largest effects would
occur in younger infants and in the poorest households, analyses of the effects
on growth were conducted by age group at baseline (<6 months and 6-12 months)
and by socioeconomic status (SES) category.
The longitudinal cohort was studied in 1999 to estimate the association
between anemia rates and exposure to Progresa during the previous year. At
the 1999 survey, the intervention communities had been receiving Progresa
for 1 year, whereas the crossover intervention communities had not. In 2000,
the cohort was also studied to estimate the differential effects of 1-year
exposure (crossover intervention communities) and 2-year exposure (intervention
communities) on anemia rates.
Written consent for participation was obtained from the mother or self-identified
decision maker in each household. The project was approved by the Human Subjects
and Ethics Committee of the National Institute of Public Health, Mexico.
Data Collection and Processing
Weight and recumbent length (in children <2 years) and standing height
(in children 2-4 years) were obtained by using standard anthropometric methods.29 Weight was measured to the nearest 10 g with an electronic
scale (model 1583; Tanita, Tokyo, Japan); length (to the nearest millimeter),
with a locally made measuring board of 1.3 m; and standing height, with a
stadiometer with the capacity to measure up to 2 m and a precision of 1 mm
(model E-1; Dyna-Top, Mexico City, Mexico). The measurements were obtained
by anthropometrists who were trained to take all measurements by using standard
techniques.30 The child's birth date was reported
by the mother and corroborated in a large proportion of children by using
birth certificates or vaccination cards, which are considered reliable documents
in Mexico. Length and weight data were transformed to z scores.2 Anthropometrists were instructed
to refer children with clinically severe malnutrition to the health centers
for treatment and to advise supervisors to follow up at the health centers.
However, no cases were reported.
A blood sample was obtained in the 1999 and 2000 surveys, when all children
were aged 12 months or older, by means of digital capillary puncture. Blood
samples were not obtained in 1998, when children were younger than 12 months,
because parents were reluctant to permit drawing blood (even the small amounts
of capillary blood required) from infants. The concentration of hemoglobin
in the blood sample was measured with a portable photoreflectometer (Hemocue,
Angelholm, Sweden). The photometer was calibrated by using the international
hemoglobin standard,31 according to the recommendations
of the International Committee of Standardization in Hematology. In 1999,
28 children with hemoglobin values lower than 9 g/dL, indicative of severe
anemia, were provided ferrous sulfate treatment and were excluded from the
analysis. For children living in altitudes above 1000 m, the hemoglobin values
were adjusted with an equation developed for Mexican populations32 to
obtain predicted hemoglobin at sea level. Anemia was defined as serum hemoglobin
concentrations lower than 11 g/dL at sea level.
Information about the frequency of papilla consumption
was obtained in the 1999 survey by interviewing the mother or the person who
fed the child. The person interviewed was asked about the usual frequency
of papilla consumption during a typical week. Children
were classified as having high supplement consumption when the mother or informant
reported consumption for 4 or more days per week.
In 1999 and 2000, socioeconomic information was collected in each household.
Information included household characteristics (construction materials used
for floors, walls, and ceilings), possession of household goods (radio, television,
VCR, telephone, refrigerator, laundry machine, and hot water heater), and
household services (water and sanitation facilities and type of fuel used
for cooking). Information was collected by combining interview with the mother
or the children's caregiver and direct observation. In 1999 and 2000, SES
scores were generated by using the first component obtained by principal components
analysis.33 Only variables with factor loadings
greater than 0.5 were retained in the model. The first component explained
49% of the total variability in 1999 and 45% in 2000. The resulting SES is
a normalized variable. Higher values of the SES correspond to better household
conditions.
Baseline characteristics of the study children, their mothers, and SES
were compared by age group between intervention groups by using a 2-tailed
test. For continuous variables, random-intercept linear models were used to
take into account the clustering effects within communities. For categorical
dependent variables, generalized estimating equations (GEE) were used to account
for clustering.
All outcomes were analyzed with the following independent variables:
SES group (below vs at or above the 50th percentile), intervention or crossover
intervention group, and infants younger than 6 months in 1998 vs infants aged
6 to 12 months, plus the 2-way and 3-way interactions. To increase statistical
power, these analyses were covaried on age as a continuous variable because
age was always statistically significant, even when age group was included.
These equations were simplified to exclude interactions if P>.10 and main effects if P>.05. Sex did not
increase power and was not included. Potential community confounders at randomization
and those that might occur thereafter were taken into account by the randomization
of the interventions to the communities.34
The increment in height from baseline in 1998 to 2000 was modeled as
a random-intercept linear model35 to take into
account that the interventions were randomized to the community and not to
the individual children. This model also takes into account clustering, meaning
children in the same community will be more similar to each other than children
across communities. Length in 1998 (centimeters) was included as a covariate
to take into account regression to the mean of the increment. Data from 1999
were not used because of the large proportion of missing data and because
1 year of intervention was considered insufficient to find effects on linear
growth.
High ingestion of papilla in 1999 (≥4 days/week)
was also investigated because it is an intermediate outcome necessary to achieve
impacts on hemoglobin concentration and growth. It was modeled with the same
main and interactive effects as was growth in height by using a GEE model.
Anemia rates for 1999 and 2000 were not related to SES or age groups.
We expected that the children who received Progresa in the first year would
have lower rates of anemia in 1999 than those children who had not received
Progresa and that the anemia rates in the latter group would decrease from
1999-2000. Therefore, we tested for the difference in anemia rates between
intervention groups in 1999. We also tested for the greater improvement in
anemia rates in the crossover intervention group as they caught up in response
to Progresa in the second year by testing the interaction between intervention
groups and year with GEE, which takes into account clustering within community
but not the positive correlation of anemia between 1999 and 2000.
Statistical significance was declared for 2-tailed tests at P<.05. Statistical analyses were done using PC-STATA (release Intercooled
Stata 8.2, Stata Corp, College Station, Tex) and PC-SAS (release 8.02, SAS
Institute Inc, Cary, NC).
At baseline, 795 children aged 12 months or younger (461 in the intervention
group and 334 in the crossover intervention group) were studied. At the end
of the study, in 2000, 650 of these children were assessed. Attrition rates
were 26% among infants aged 5 months or younger and 10% among those aged 6
to 12 months, with little difference between the intervention and crossover
intervention groups (Figure 1).
In 1999, the proportion of children who were not assessed was much greater.
There was a 45% loss of measurement among those aged 5 months or younger and
a 32% loss among those aged 6 to 12 months, which was due to an unusually
long and severe rainy season that resulted in a high rate of temporal emigration
from a large number of communities because of damage caused by severe flooding.
Table 1 shows descriptive
statistics of the main outcomes studied in the panel of infants who were followed
up longitudinally for 2 years. Data are presented by age categories at baseline.
Blood samples were not obtained for infants, and socioeconomic information
is unavailable for 1998; therefore, hemoglobin and SES data were unavailable
at baseline. No significant differences were found in the comparison of the
main outcomes between intervention and crossover intervention infants.
Papilla intake information was available for
341 children, 72% of the total number of children for which information is
available in 1999. Fifty-seven percent (n = 122) of the children in the intervention
communities consumed papilla regularly (≥4 days/week)
in 1999 at the end of the first year of intervention. As a result of program
leakages, there was some regular ingestion of papilla in
the crossover intervention communities during 1999 (13 children; 9.7%). Among
the intervention group infants in 1999, 58.0% of older infants (6-12 months
of age) and 55.8% of infants younger than 6 months consumed papilla at least 4 days a week (P = .09).
Adjusting for age, a higher percentage of the poorest infants ingested papilla at least 4 days a week compared with the less-poor
infants (61.0% vs 52.9%; P = .22) in the intervention
group in 1999. During the second year, the amount of papilla ingested was not significantly different between the intervention
group and the crossover intervention group (P = .18).
Figure 2 shows the adjusted
increments in height by age, length, SES, and intervention groups. The expected
growth from the WHO reference standards is plotted in the figure for comparison.2 The triple interaction between age, SES, and intervention
groups was significant (P = .04), showing that the
effect of exposure to Progresa was different among the 4 subgroups (age at
baseline and SES subgroups). This triple interaction was due to the greater
effect (P = .046) on infants younger than 6 months
at baseline and who lived in the poorest households. Among these infants,
adjusted height was greater by 1.1 cm (26.4 cm in the intervention group vs
25.3 cm in the crossover intervention group).
Mean hemoglobin values were higher in the intervention group (11.12
g/dL; 95% confidence interval [CI], 10.9-11.3 g/dL) after 1 year in the Progresa
program relative to the crossover intervention group (10.75 g/dL; 95% CI,
10.5-11.0 g/dL) who had not received Progresa benefits that year (P = .01). In 2000, the difference between intervention groups was no
longer significant (P = .26) after both groups had
been in the program during the previous year.
Figure 3 shows that the prevalence
of age-adjusted anemia in 1999 was higher in the crossover intervention group
(54.9%) than in the intervention group (44.3%) (P =
.03), whereas in 2000 the difference was not significant (23.0% vs 25.8%,
respectively) (P = .40). However, the change in the
difference between the intervention or crossover intervention group that occurred
from 1999 to 2000 was significant (P = .002) for
the interaction term, even though the method to estimate the interaction underestimates
its significance.
This randomized effectiveness study showed that Progresa, a large-scale
government program that includes a nutrition intervention, is associated with
better growth in height and lower rates of anemia in low-income rural Mexican
infants.
Benefits in height were found in infants with the highest response potential:
the poorest (below the 50th percentile of SES distribution) and younger infants
(<6 months at baseline). Among this group, infants in the intervention
group had a 2-year increment in height of 1.1 cm relative to the crossover
intervention group (Figure 2). This
statistically significant difference is also biologically important; it is
one fourth of the deficit in growth of the youngest poorest crossover intervention
infants relative to their genetic potential, as estimated from the WHO reference
standards2 for children of the same age. Growth
benefits occurred as expected in the subgroup with the highest response potential,
reinforcing the biological plausibility of the results.36
This study has several limitations. One concern was a possible selection
bias caused by loss to follow-up, particularly in 1999 when rates of loss
to follow-up were high. However, no differences in any of the baseline variables
were found between children lost to follow-up and those who were not for either
intervention group or year (1999 and 2000).
The results may underestimate the real effects of Progresa interventions
because the group that was originally allocated to be the control was included
in the Progresa program 1 year earlier than planned. However, the crossover
intervention group's exposure time was shorter than that of the intervention
group and introduced during an age period that has lower potential for growth
response. Had the crossover intervention group remained a control group during
the first 2 years of the study, the height increments of this group might
have been smaller, and therefore the difference between intervention groups
might have been greater. In addition, about 10% of infants in the crossover
intervention group were receiving papilla throughout
the first year of the evaluation because of program leakage. These children
benefited from the program during the same period of high growth-response
potential as the intervention group, thus diminishing the difference between
the groups.
The intervention group had better growth in height than the crossover
intervention group, but it was not sufficient to achieve adequate growth.
At the end of the evaluation, the poorest young infant group benefited the
most from Progresa. They were aged 24 to 29 months and 1.1 cm taller, but
they still had a height deficit of 3.4 cm relative to the height-for-age distribution
recommended by WHO2 for this age period. This
deficit is equivalent to about 1 SD.
In a literature review of effectiveness studies aimed at improving the
nutritional status and growth of infants,14 only
3 studies were identified as having height or height-for-age z score as an outcome. At baseline, infants in the studies were aged
6 to 12 months. The effect documented in our study (z score,
0.3) is similar to that of 2 of the studies reviewed (z score, 0.3-0.4) but smaller than that reported in the third study
(z score, 0.87). However, baseline mean height-for-age z score was high in Progresa children (−1.06) relative
to the 3 studies reviewed (−1.8 to −1.6). Therefore, smaller effects
would have been expected in Progresa children.
In 1999, after 1 year of program implementation, the rate of anemia
in the intervention group was significantly lower than that in the crossover
intervention group that had not yet received the benefits of Progresa (Figure 3). In the last follow-up survey,
at the end of 2000, the difference in anemia rates observed in 1999 had disappeared.
Our interpretation of this finding is that Progresa improved the hemoglobin
concentration in the crossover intervention group, which received the benefits
of the program in the second year ending in 2000. Anemia in both groups in
2000 was below the national average prevalence of 34.0% for poor rural children
of the same age groups,17 indicating that both
groups benefited. The difference in 1999 in anemia rates is important but
is nevertheless insufficient, considering that the anemia rates remained high,
at more than 44%, in the intervention group that benefited from Progresa during
the first year. In addition, although the prevalence of anemia declined to
similar rates in both groups after year 2, approximately one quarter of children
who received Progresa had anemia.
It is unlikely that the high anemia rates are due to nonnutritional
causes of anemia such as parasitic infections because malaria and hookworm,
2 of the parasites that contribute most to anemia, are uncommon in Mexico.
For example, only 786 new cases of malaria were reported in Mexico in 2002,37 and a literature review of studies conducted between
1981 and 1992 in high-risk populations in Mexico reported a pooled general
prevalence of hookworm infections of 1.8%.38
A possible reason for the inadequate effects of the nutritional supplement
may be the fact that reduced iron, which is not absorbed well, was used for
the fortification of papilla (J.A.R. et al, unpublished
data).39 As a result of this evaluation, a
number of studies are now being undertaken to identify a more bioavailable
iron compound for use in the fortification of papilla.
In addition to the intake of papilla, other
Progresa interventions that may have contributed to the nutritional impact
should be considered, including other dietary improvements associated with
cash transfers and maternal nutrition education,REFLINK IDREF="REF-JOC40268-40">
lowered diarrhea rates because of better sanitation and quality of housing
through cash transfers, and an increased use of primary health care.20,41 The program has expanded to urban
areas, and we are evaluating its impact on dietary intake and health care
use.
This is the first evaluation of the nutritional impact of a large-scale
program in Mexico and one of few in developing countries,14 and
results are being used for program improvement. This experimental study showed
improvements in height increments and anemia rates associated with exposure
to Progresa, a program with a nutrition component, which covered about 2.6
million families by the end of the evaluation period and that today reaches
4.5 million families. The randomized design of the study allowed us to attribute
the biologically important differences between intervention groups to the
effect of the program. This evaluation is important because it is one of the
few well-controlled effectiveness evaluations of large-scale programs that
showed positive effects on nutrition and that also promises to support informed
program improvements.
1.Black RE, Morris SS, Bryce J. Where and why are 10 million children dying every year?
Lancet.2003;361:2226-2234.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12842379&dopt=Abstract
Google Scholar 2.World Health Organization. Measurement of Nutritional Impact. Geneva, Switzerland: World Health Organization; 1979.
3.Pelletier DL, Frongillo Jr EA, Habicht J-P. Epidemiologic evidence for a potentiating effect of malnutrition on
child mortality.
Am J Public Health.1993;83:1130-1133.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=8342721&dopt=Abstract
Google Scholar 4.Fishman S, Caulfield LE, de Onis M.
et al. Underweight. In: Essati M, Lopez AD, Rodgers A, Murray CJL, ed. Comparative Quantification of Health Risks: Global and Regional
Burden of Disease Attributable to Selected Risk Factors. Geneva, Switzerland: World Health Organization. In press.
5.Martorell R. Results and implications of the INCAP follow-up study.
J Nutr.1995;125:1127S-1138S.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=7536833&dopt=Abstract
Google Scholar 6.United Nations Administrative Committee on Coordination. Fourth Report on the World Nutrition Situation. Geneva, Switzerland: United Nations Administrative Committee on Coordination/Sub-Committee
on Nutrition; 2000.
7.Habicht JP, Butz WP. Measurement of health and nutrition effects of large scale nutrition
intervention projects. In: Klein RE, ed. Evaluating the
Impact of Nutritional and Health Programs. New York, NY: Plenum Press; 1979.
8.Gwatkin DR, Wilcox JR, Wray JD. The policy implications of field experiments in primary health and
nutrition care.
Soc Sci Med.1980;14C:121-128.Google Scholar 9.Caulfield L, Huffman SL, Piwoz EG. Interventions to improve intake of complementary foods by infants 6
to 12 months of age in developing countries: impact on growth and on the prevalence
of malnutrition and potential contribution to child survival.
Food Nutr Bull.1999;20:183-200.Google Scholar 10.Rivera JA, Hotz C, González-Cossío T, Neufeld L, García-Guerra A. The effect of micronutrient deficiencies on child growth: a review
of results from community-based supplementation trials.
J Nutr.2003;133:4010S-4020S.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=14672304&dopt=Abstract
Google Scholar 11.Jones G, Steketee RW, Black RE, Bhutta ZA, Morris SS, Bellagio CSSG. How many child deaths can we prevent this year?
Lancet.2003;362:65-71.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12853204&dopt=Abstract
Google Scholar 12.Beaton GH, Ghassemi H. Supplementary feeding programs for young children in developing countries.
Am J Clin Nutr.1982;35(4 suppl):863-916.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=7039297&dopt=Abstract
Google Scholar 13.Horweg J. Impact evaluation of child nutrition programmes.
Food Policy.1988;13:199-207.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12341717&dopt=Abstract
Google Scholar 14.Caulfield L, Piwoz EG, Huffman SL. Improved child diet and growth: efficacy and effectiveness studies.
In: Nutrition and Growth: Nestlé Nutrition Workshop
Series Pediatric Program. New York, NY: Nestle; 2001;47:375-403.
15.Rivera JA, Lutter Ch. The potential role of processed complementary food in Latin America.
In: Nutrition and Growth: Nestlé Nutrition Workshop
Series Pediatric Program. New York, NY: Nestle; 2001;47:281-303.
16.Barquera S, Rivera-Dommarco JA, Gasca-García A. Políticas y programas de alimentación y nutrición
en México.
Salud Pública Mex.2001;43:464-477.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11765723&dopt=Abstract
Google Scholar 17.Rivera JA, Sepúlveda-Amor J. Conclusions from the Mexican National Nutrition Survey 1999: translating
results into nutrition policy.
Salud Pública Mex.2003;45:S565-S575.Google Scholar 18.Sen A. Development as Freedom. Oxford, England: Oxford University Press; 1999.
19.Skoufias E, Davis B, Behrman J.
An Evaluation of the Selection of Beneficiary Households
in PROGRESA: Final Report. 1999.
International Food Policy Research Institute Web site. Available
at: http://www.ifpri.org/themes/progresa/targeting.htm. Accessed
May 4, 2004. 20.Gertler P, Rivera J, Levy S, Sepúlveda-Amor J. Mexico's PROGRESA: using a poverty alleviation program as an incentive
for poor families to invest in child health.
Lancet.Submitted.Google Scholar 21.Schroeder DG, Martorell R, Rivera JA, Ruel MT, Habicht JP. Age differences in the impact of nutritional supplementation on growth.
J Nutr.1995;125:1051S-1059S.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=7722707&dopt=Abstract
Google Scholar 22.Rivera J, González-Cossío T, Flores M, Hernández M, Lezana MA, Sepúlveda-Amor J. Déficit de talla y emaciación en menores de cinco años
en distintas regiones y estratos en México.
Salud Pública Mex.1995;37:95-107.Google Scholar 23.Rivera J, Bourges H, Arroyo P.
et al. Deficiencia de Micronutrimentos. Diez problemas relevantes de salud pública en México. Instituto Nacional Salud Pública. 1999:7-57.
24.Rosado JL, Rivera J, López G, Solano L. Development, production, and quality control of nutritional supplements
for a national supplementation program in Mexico.
Food Nutr Bull.2000;21:30-34.Google Scholar 25.Rothman KJ, Greenland S. Modern Epidemiology. 2nd ed. Baltimore, Md: Lippincott-Raven; 1998:144.
27.Lohr SL. Sampling: Design and Analysis. Pacific Grove, Calif: Duxbury Press; 1999:241.
28.Murray DM. Design and Analysis of Group-Randomized Trials. New York, NY: Oxford University Press; 1998:364.
29.Lohman T, Roche A, Martorell R. Anthropometric Standardization Reference Manual. Champaign, Ill: Human Kinetics Publishers; 1988.
30.Habicht JP. Estandarización de métodos epidemiológicos cuantitativos
sobre el terreno (standardization of anthropometric methods in the field).
PAHO Bull.1974;76:375-384.Google Scholar 31.Zwart A, van Assendelft OW, Bull BS, England JM, Lewis SM, Zijlstra WG. Recommendations for reference method for haemoglobinometry in human
blood (ICSH standard 1995) and specifications for international haemiglobinocyanide
standard.
J Clin Pathol.1996;49:271-274.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=8655699&dopt=Abstract
Google Scholar 32.Ruiz-Arguelles GJ, Llorente-Peters A. Predictive equations for normal red cell values at altitudes from sea
level to 2,670 meters.
Rev Invest Clin.1981;33:191-193.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=7291765&dopt=Abstract
Google Scholar 33.Mardia KV, Kent JT, Bibby JM. Multivariate Analysis. London, England: Academic Press; 2000:213-216.
34.Hinkelman K, Kempthorne O. Design and Analysis of Experiments. New York, NY: Wiley; 1994.
35.Raudenbush SW, Bryck AS. Hierarchical Linear Models. 2nd ed. Thousand Oaks, Calif: Sage Publications; 2002:26-27.
36.Habicht J-P, Victora CG, Vaughan JP. Evaluation designs for adequacy, plausibility and probability of public
health programme performance and impact.
Int J Epidemiol.1999;28:10-18.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=10195658&dopt=Abstract
Google Scholar 37.Sistema Nacional de Vigilancia Epidemiológica. Epidemiologia. 2003:14.
38.Tay J, Ruiz A, Sanchez-Vega J, Romero-Cabello R, Becerril MA. Intestinal helminthiasis in the Mexican Republic.
Bol Chil Parasitol.1995;50:10-16.Google Scholar 39.Turner L. Monterrey Workshop summary: evaluating the usefulness of elemental
iron powders.
Nutr Rev.2002;60:S16-S17.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12141593&dopt=Abstract
Google Scholar 40.Hoddinott J, Skoufias E. The Impact of PROGRESA on Food Consumption: Discussion
Paper 150. Washington, DC: International Food Policy Research Institute; 2003.