Health in Mexico: Perceptions, Knowledge and Obesity*
Susan W. Parker, Spectron Desarrollo/CIDE
Luis N. Rubalcava, Spectron Desarrollo/CIDE
Graciela M. Teruel, Universidad Iberoamericana
* This research was financed by the InterAmerican Development Bank under the project
title: Understanding Quality of Life in LAC: A multidimensional approach. We thank
Maria Jose Gonzalez for excellent research assistance.
1
1. Introduction:
This paper focuses on health in Mexico. Health in Mexico is undergoing major
transitions, with a continued rise in life expectancy accompanied by a shift away from
infectious diseases as the major cause of death. Obesity is one of the main risk factors
associated with the principal causes of death (diabetes and cardiovascular diseases) and
studies have suggested that elevated body mass index (BMI) represent the largest cause
of premature death in Mexico. (Secretaria de Salud, 2004).
This paper has three main objectives. First, we use the MxFLS survey to provide a
general multi-dimensional description of health in Mexico between 2002 and 2005.
Secondly, we take advantage of anthropometric information in the MxFLS to analyze
the extent to which people are knowledgeable of their health conditions. We analyze for
variables including weight, height, blood pressure and diabetes, the extent to which
what individuals report correlates with their actual or measures and analyze the extent to
which reporting errors are due to lack of knowledge about true health conditions or
embarrassment in reporting, such as in the case of weight.
Finally, we provide evidence on the problem of obesity in Mexico, a country with one
of the highest levels of obesity in the world. We document changes in obesity over
time, its persistence and finally analyze the extent to which change in obesity are
correlated to changes in mental health.
2
2. Previous Literature:
The relationship between health perception and measured health has a voluminous
literature (see Sadana, 2001 for a review and Kyffer et al. 2004 and Rahman and Barski,
2003 for some recent examples) with many studies comparing self-reported morbidity
indicators with indicators based on health examinations by health workers and other
studies analyzing how self-reported health predicts health problems later in life.1
The issues of differences in reported versus actual health have also begun to be studied
in the economics literature, where the emphasis has been on how mis-measurement of
health may lead to misleading conclusions in economic studies of health. For instance,
Strauss and Thomas, 1996 analyze reported height versus measured height in Zimbabwe
and show how income appears to affect these differences, with income predicting lower
differences between reported and measured height.
Thomas and Frankenburg, 2001 provide an excellent theoretical framework for
modeling self-reported health and objective or measured health, where measured health
differs from self-reported health because of reporting effects, and measured health may
differ from actual health status, which is unobserved. They use the Indonesian Family
1 Deaton,
2006 analyzes the relationship of life and health satisfaction to income, age
and life expectancy for 132 countries using the Gallup 2006 survey. He shows that
while average happiness is strongly related to per capita income, this is not the case for
self-reported health where there is little relationship to per-capita income. These results
lead Deaton to question the usefulness of self-reported health as a welfare indicator for
use in international comparisons. This lack of relationship however might partly reflect
a lack of knowledge of true health conditions, or the lack of influence of true health
conditions on health perceptions.
3
Life Survey to carry out several exercises relating socio-economic status, measured
health and self-reported health, based on the indicators of height and weight (indicators
for which they have both measured and self-reported information). In particular, they
analyze the difference in reported versus actual weight and height, finding that men tend
to overstate their weight whereas women tend to understate their weight. Nevertheless,
the extent of under or over reporting interacts in a complex way with age and education.
They also analyze the impact of measured health on self-reported health, finding some
significant relationships of hemoglobin and blood pressure on reported health status.
Our study builds on Thomas and Frankenburg by studying similar issues in the Mexican
context and focusing on the extent to which reporting biases reflect lack of knowledge
of true health conditions, which may be particularly important for the chronic diseases
of diabetes and hypertension, where patient lack of knowledge is common. In particular,
errors in reporting may derive from at least two explanations. First, for some indicators,
individuals may feel embarrassment at reporting the truth (such as for weight).
Alternatively, reporting errors may reflect lack of knowledge about true health
conditions, for instance in the case of blood pressure and diabetes, where
embarrassment costs might be lower than say for weight and misreporting is more likely
to be due to lack of knowledge. Of course there may also be interviewer errors and
coding errors, as is always the case in fieldwork. Our analysis on reporting errors then is
useful not only as informative on the accuracy of self-reported health information but
also on the extent to which individuals may have undiagnosed diseases.
Obesity
4
Studies on obesity in Mexico have mainly restricted themselves to measuring levels of
obesity at a national level and within population sub-groups. There are no studies we
could find which provide direct evidence on the causes of high and increasing obesity
levels in Mexico. Existing previous studies have shown that Mexico has an extremely
high incidence of obesity with urban areas reporting slightly higher rates than very rural
areas (Fernald et al. 2004) and women reporting significantly higher rates of obesity
than men (30% of Mexican women are obese (BMI >30) versus 20% of men in 2000).
With respect to trends, Garnier et al., 2005 analyze changes in obesity of Mexican
women between 1987 and 1999 and find the proportion of women who are overweight
or obese (BMI >=25) increases over this twelve year period from 28.4 percent to 54.8, a
rapid increase by any standard.
With respect to obesity and mental health, there is a fairly large medical and
psychological literature studying the relationship between obesity and mental health,
although we have found no studies on Mexico in this area. An obvious issue is
causality here, e.g. does being overweight cause mental health problems or does the
prevalence of mental health problems tend to cause individuals to gain or lose weight?
The available studies tend to emphasize the potential effect of mental health on
becoming at risk for obesity, focusing on the prevalence of psychiatric disorders among
the obese population. McElroy et al. review the available literature on obesity and
mental health disorders and conclude that “children and adolescents with major
depressive disorder may be at increased risk for developing overweight” and that
“patients with bipolar disorder may have elevated rates of overweight, obesity, and
abdominal obesity.” On the other hand, Felton and Graham using data from the United
States and Russia argue that in these particular contexts, there is some evidence that the
5
causality goes from being overweight to depression. Felton and Graham also argue that
stigma affects the well-being costs (measured by say depression) of obesity providing
evidence that whites in the United States have a higher stigma associated with obesity
than blacks or Hispanics who overall tend to have higher rates of obesity.2
3. Data:
MxFLS-1 is a broad-purpose multi-topic, nationally representative survey of
individuals, households and communities. The baseline covers over 8,400 households in
150 communities across the whole Mexico. All individuals age 15 and over were
interviewed and extremely detailed information on a wide array of social, economic,
demographic and health behaviors of individuals and their families was collected. All
household members participated in an in-home physical health assessment which
measured anthropometry, hemoglobin levels and blood pressure. A Raven's (cognitive)
test was applied to all members of the household in specific age categories, and a short
mental health instrument was administered to all adults. The second wave, MxFLS-2
began fieldwork in the first semester of 2005 with a similar survey content as the first
round.
All individuals and households moving were followed up and follow up
interview rates of about 90 percent were achieved.
We now turn to the specific information available on individual health in the MxFLS.
There is an extensive health survey applied to all individuals. For individuals over age
2 Some recent research (Christakis and Fowler, 2007) emphasizes the impact of social
networks in weight gain, concluding that weight gain in friends or neighbors may be as
important to affecting one’s own weight gain as weight gain in closer relatives such as
spouses.
6
15, it is answered directly by the individual in both the 2002 and 2005 wave. There are
several modules in this health survey, including questions about their recent health
condition, emotional well-being, chronic illness and in- and out patient utilization. The
health conditions module include questions ranging from how you would rate your
health, and how would you rate your health in comparison with others of the same age
gender, to incidences of illness, previous accidents, ability to carry out a number of
different activities of daily living (for those above age 50), and morbidity data, such as
coughs, headache, diarrhea and fever. The emotional well-being section contains 21
questions asking about feelings in the past four weeks relating to loneliness, insecurity,
sleeplessness, sadness. The chronic illness covers diagnosis and treatment of diabetes,
hypertension, heart disease, cancer etc.
The utilization module covers in-patient
treatment over the past year as well as out-patient treatment in the past four weeks. The
MXFLS includes biomarker data on hemoglobin levels, blood pressure, waist to hip
ratios, height and weight. In 2005, additionally glucose levels, total cholesterol and the
collection of dried blood spots were collected for adults above 15.
4. Health Indicators in Mexico 2002-2005
One of the major advantages of the MxFLS is the point that it is longitudinal,
nevertheless, most longitudinal data sets have high attrition which may not be random
with respect to the variables of interest. A specific objective of the MxFLS is to
minimize attrition, Table 1 provides information on the sample size in 2002 and 2005
and what percentage of the original 2002 sample have data for 2005. The table shows
between 2002 and 2005 a low attrition rate of between 10 and 15% was achieved,
depending on the category of health indicators. Note this data does not yet include
information for those who migrated from Mexico to the United States between 2002 and
7
2005 although the data does include information on individuals who migrated within
Mexico. (The international migrants data has not yet been coded but when available will
result in further reductions in attrition levels for many of the variables reported here).
Table 2a provides a summary of health indicators found in the MxFLS in 2002 and
2005 and the difference over time for individuals between 20 and 60 years. With respect
to the anthropometric and biomarker measures, there is a notable high degree of obesity
with about two thirds of the population either overweight or obese in 2002. The
proportion of those overweight/obese increases by about 2 percentage points between
2002 and 2005. With respect to blood pressure, there are some important shifts over
time, whereas the population with hypertension (above 140/90) falls significantly
overtime, the proportion with high blood pressure (above 120/80) increases
substantially.
Glucose levels are measured only in 2005. To identify the population having diabetes in
our sample, we use a cutoff level of a blood sugar reading of 126 mg/dL. Nevertheless,
an important issue is whether the glucose level was measured after an eight hour fast as
recent meals tend to increase blood sugar levels. The MxFLS fortunately attempted to
insure that readings were done after an overnight fast and provides information within
the data when this was not possible. Thus, we disaggregate glucose levels by whether
the reading was done after an (at least) 8 hour fasting. Of those who fasted (about two
thirds), about 10% of the population have levels which are considered to be high and
symptomatic of diabetes. For those who did not fast, we used a higher reading to
measure the incidence of diabetes (above 186mg/dL) about 5 percent have diabetes.
About 5 percent of the adult population have high levels of cholesterol.
8
The MxFLS also has information on symptoms, e.g. the proportion of the population
reporting flu, coughs respiratory problems, diarrhea headache during the previous four
weeks. What is notable about this data are the trends over time. For all of the symptoms,
a large decrease is reported for the same individuals (who are three years older in 2005)
over time. This data is consistent with some important improvements in health overtime
in the Mexican population.
Next, we report activities of daily living, a set of questions which is only applied to
those age 50 and over. As might be expected, the proportion of those reporting ability to
carry out the basic activities of daily living decreases over time, reflecting the aging
process.
Finally, there is also a section of the MxFls on self-reported chronic diseases. The
proportion of the population reporting they have diabetes is about 6 percent and
constant overtime, this is clearly much lower than the estimates based on glucose
measurements reported above. Two percent of the population report heart disease and
one percent report cancer. About 16 percent of adults sought medical attention in the
past 4 weeks in 2002, a proportion which decreases to 11 percent by 2005, a significant
decrease over time.
Table 2b explores the sources of the fall in health clinic consultations, providing the
total number of visits by type of provider. Visits in the MxFLS by all of the major
public and private providers of consultations fall substantially between 2002 and 2005.
IMSS and ISSSTE visits by the MxFLS population fall by about 35% and visits to
private doctors fall also by 36%. There is also a large fall in visits to the Red Cross.
9
The only significant provider where visits increase overtime is in drugstores. Further
analysis is required to identify why visits fall by such a large proportion, and whether
visits fall primarily in preventative or curative care.
Table 3 provides the same health variables by gender and rural/urban residence in 2002.
With respect to anthropometrics, the proportion of women who are overweight/obese is
similar in rural and urban areas, at nearly 70 percent. Men have slightly lower levels of
obesity than women, particularly in rural areas although even there the proportion of
those overweight/obese is above 60 percent. Blood pressure levels are higher for men,
both in rural and urban areas.
Looking at health symptoms in the previous four weeks, it is notable that levels of
symptoms such as cough, flu, nausea, body ache, and diarrhea show much higher levels
in rural areas than in urban areas. In rural areas, for many indicators men and women
report similar levels of these symptoms, in urban areas women report a greater degree of
health symptoms than men. Most health indicators/symptoms suggest worse health for
those living in rural areas in both years of the MxFLS, as one might expect given higher
poverty rates and lower availability of health services in rural areas with respect to
urban areas.
With respect to the activities of daily living (ADL’s), overall the rural population over
age 50 reports greater difficulty overtime in carrying out strenuous activities such as
walking 5 kilometers or climbing stairs, as one would expect from a population of older
adults. Women in both rural and urban areas report far greater problems in carrying out
ADL’s than men. Women are much more likely to have visited a health clinic in the
10
previous four weeks than men, with similar levels in rural and urban areas (about 25%
of women attended a health clinic versus 11% of men.)
Table 4 repeats the previous table but reports changes in health between 2005 and 2002
for the rural and urban population and for men and women. On average, in rural areas,
men and women increase weight in about 1 kilogram over the three year period whereas
in urban areas the increase is about three quarters of a kilogram.
Turning to health symptoms suffered during the previous four weeks, there are
reductions among both men and women in both rural and urban areas. The sharpest
reductions in the incidence of health symptoms occurs in rural areas for both men and
women. With respect to activities of daily living, the largest reductions in the proportion
able to carry out activities of daily living are in rural areas (Table 4).
In summary, in this section we have provided a broad description of the health of the
Mexican population as well as how health perceptions and levels are changing over the
period. There are several notable characteristics of Mexican health. First, the proportion
of adults who are obese or overweight is very large, and increases slightly over the
panel period. According to glucose tests, the proportion of the adult population with
diabetes is large although a smaller proportion of the population report they have
diabetes, implying a fraction of the population may not be aware. High blood pressure
is also quite prevalent among the adult population and increasing over time.
The news is not all bad however, there are some areas where better health as measured
by lower incidences of health problems such as coughs, flu, diarrhea, headaches and
11
fever is occurring. These positive changes occur both in urban and rural areas and
among men and women.
As might be expected, rural men and women have overall worse health as measured by
the above symptoms and they also report more difficulty with the activities of daily
living than those in urban areas. Obesity levels remain somewhat lower for men in rural
areas than in urban areas, although female obesity levels are quite similar in rural and
urban areas.
5. Reporting errors and knowledge of health conditions.
For several anthropometric and biomarker indicators we have information on actual
measured health status and also on what individuals report. This information allows us
to carry out an analysis of reporting biases. The health information in which MxFLS
provides both objective and self-reported health outcomes are: height and weight,
hypertension and diabetes, as defined by measured glucose levels after fasting.
Differences between actual health measures and reported health measures might exist
because of lack of information of true health conditions. This is likely to be particularly
relevant for the cases of diabetes and high blood pressure, where even in developed
countries such as the United States, a relatively large fraction of those having diabetes
are not aware they have the problem. There may also be a reporting bias, for instance in
the case of weight, some individuals may feel embarrassed about reporting their true
weight if they are overweight.
Similarly individuals may overstate their height.
Depending on the variable then, differences between reported and actual health may
12
reflect either lack of information or embarrassment/psychic costs associated with
reporting the truth. Our multivariate analysis will provide some intuition on the relative
existence of “lack of knowledge” reporting errors versus embarrassment or psychic
costs. For instance, if lack of knowledge is the primary reason self-reports differ from
measured health, one might expect individuals with more education, or individuals who
have recently visited health clinics to have more knowledge about their own health
status.
For weight and height we will look at the difference between the measured and selfreported anthropometric outcome. We compare if the person reports suffering from
diabetes and hypertension with actual measures of glucose and blood pressure levels
taken by a health worker during the MxFLS interview, respectively. We also study how
these reporting biases differ by socioeconomic status, age, gender, education and
indigenous status.
There is a complication for the analysis of reporting errors with respect to diabetes and
hypertension. In particular, some individuals who have diabetes are likely to be aware
of this and seek treatment, it is possible that some individuals with diabetes who are
being treated may achieve sufficiently low levels so as to be defined as not having
diabetes in our analysis. We will discuss how we approach this measurement issue in
more detail below.
Tables 5 and 6 provide some general descriptions of the extent to which people report
their health status correctly. Table 5 reports data with respect to weight and height. For
weight, we define accurate reporting to be when the difference between measured
13
weight and self-reported weight is less than one half of one standard deviation (about 7
kilos). About 65 percent of individuals accurately report their weight. However there are
important differences by gender, women overall have a higher proportion reporting
accurately their weight, however when they do not report accurately they tend to under
report their weight more than they over-report. Men are less likely to accurately report
their weight but tend to under and over report their weight in similar proportions. This
evidence is both suggestive that women have more information about their weight
(perhaps because they frequent health clinics more often than men as shown in the
descriptive) but also may suffer from embarrassment at reporting their true weight.
Table 5 also shows reporting errors for height. Here only about half the population
accurately reports their height and both men and women have a high tendency to overreport their height. This may reflect lack of information or “shrinking” with aging or it
may also reflect a perceived social status to being taller (one might have expected more
over-reporting of height by men but this does not seem to be the case).
Table 6 divides the population into four groups to address reporting differences in
hypertension and diabetes: those who do not have the disease and correctly report this,
those who do have the disease and are aware they have the disease, those who do not
have the health problem but report having it and those who have the disease but are
unaware. What is striking is focusing on the population who do in fact have the illness.
In the case of hypertension, about 3 percent of the population have the illness and
correctly report they have it. However, 13% have the disease and report incorrectly
they do not have the illness. If this in fact represents lack of knowledge of the disease
as opposed to embarrassment at reporting, it would suggest that about 80 percent of
14
those with hypertension are unaware they have this health problem. Similar findings
occur for diabetes. Here we again distinguish between those who fasted before the
glucose measure was taken versus those who did not. For those fasting, about four
percent of the population have the health problem and are aware of having it whereas
7% of the population has diabetes and is unaware.
We now turn to characteristics of the population with reporting errors versus those
without. For height and weight, we define the reporting error as reported minus
measured weight/height for all persons. For hypertension and diabetes, however, we
focus only on the population with the disease and analyze who is aware of the disease
and who is not.
As mentioned above, one issue to consider is how to treat those who report having
diabetes/hypertension but the biomarker analysis showed they did not, e.g. their blood
sugar reading was below 126 mg/dL and their blood pressure was over 140/90.
Treatment of both diseases can lead to normal readings, in which case such individuals
should be treated as having diabetes/hypertension and being aware of the disease.
However, these individuals could also simply have reported incorrectly they had
diabetes in which case they should not be considered in the population that has diabetes.
We carry out the underreporting analysis both ways, by including and then excluding
these individuals from the sample with diabetes.
Tables 7 through 10 present regression analysis of reporting errors. Turning to Table 7,
which analyzes reporting errors of weight, we use two particular specifications. The
first is the simple difference, reported-measured weight, the second is a probit analysis
15
for whether weight was accurately reported, as defined by being within half a standard
deviation. Overall, there are few significant determinants of reporting errors. Men are
less likely to accurately report weight as shown in the descriptive analysis. Seeing a
doctor in the past three months is associated with an increased probability of reporting
weight correctly, suggesting that individuals may acquire information about their weight
from health appointments. The most significant variables however affecting accurate
reporting of weight are the actual level of weight. Obese individuals tend to underreport their weight relative to those of normal weight by about 2.5 kilograms and are
less likely to accurately report their weight by about 6 percentage points. Very
overweight individuals may feel embarrassed to report their true weight, or may engage
in a bit of “wishful” thinking over their actual weight. Our analysis is thus suggestive of
both information effects in reporting errors and misreporting due to embarrassment
motives.
Table 8 reports a similar table for reporting errors in height. Obese individuals tend to
over-report their height and seeing a doctor during the past three months has a positive
correlation with accurately reporting weight. Other than these variables, there are few
significant predictors of reporting errors in height.
For the measures of hypertension and diabetes (Tables 9 and 10), we use the sample of
individuals with the disease, and analyze the determinants of reporting not having the
disease. We provide regressions both considering those who state they have the disease
but whose biomarkers do not reflect that as having the disease and regressions where we
exclude them from the analysis, effectively assuming they have incorrectly reported
having the disease.
16
Turning to Table 9 we study the factors associated with underreporting hypertension.
Men are much more likely than women to underreport or to be unaware they have
hypertension, again perhaps of their lack of attendance at health clinics. Younger
individuals are more likely to be unaware compared with older individuals. Having
been hospitalized and having seen a doctor in the previous three months are associated
with a reduction in the incidence of not knowing about having the disease. Individuals
with more education are more likely to be aware of having the disease. A simple
dummy for being an Oportunidades beneficiary is not significant.
Table 10 contains determinants of underreporting/being unaware of having diabetes.
Older individuals and males are less likely to be unaware of having diabetes. Obese and
overweight individuals are more likely to be unaware of having diabetes relative to
individuals with normal body weights. Having recently visited a doctor or having
recently been hospitalized reduces the likelihood of being unaware of having diabetes.
These variables are however simple correlations, the causality may simply be that
individuals who are aware they have diabetes are more likely to have recently visited a
healthcare provider.
In summary, in this section we have demonstrated the importance of reporting errors in
health and discussed their interpretation. Reporting errors in health may occur due to
lack of knowledge of true health conditions, embarrassment or other motives which
cause purposeful misreporting and interview errors. Our analysis suggests that
information and embarrassment motives are important with respect to reporting weight
and height. With respect to diabetes and hypertension, a large proportion of the
17
population with the disease reported they did not have the disease. This is clearly an
important public health problem. Our multivariate analysis predicting who was aware
of having diabetes or hypertension revealed few significant variables, including actual
weight (reducing probability of being aware), gender (males are less likely to be aware)
and age (younger individuals are more likely to be unaware).
6. Trends and transitions in obesity in Mexico
In this section, we provide a more in depth study of the problem of obesity in Mexico, a
country with one of the highest rates of obesity in the world.
We first turn to the persistentness of obesity. Table 11 divides the population into four
groups, those who were obese in 2002 and 2005, those who were not obese in either
year, and those who were not obese in 2002 but became obese in 2005 and those who
were obese in 2002 but not in 2005. Table 11 presents the distribution between the four
groups nationwide, for men and women and for urban and rural areas. There is some,
perhaps surprising, mobility in and about of obesity, although most of those obese at a
moment in time are obese continuously. Whereas about 22 percent of the population is
obese in both years, about 7 percent of the population is obese in 2002 and not obese in
2005 and about 8 percent of the population is not obese in 2002 and becomes obese by
2005. Women have higher proportions of continuous obesity than men.
The previous table suggested that about 15 percent of the population change from being
obese to not being obese over time, however given the cutoff nature of the BMI
definition (over 30), this might reflect very small changes around the cutoffs. Table 12
18
presents changes in weight and BMI for the four groups and shows actual significant
changes in weight for the groups reporting changes in obesity status overtime. The
group of individuals who was obese in 2002 and is not obese in 2005, loses on average
10 kilos between the three years while the group of individuals who become obese over
time on average gains about 11 kilos.
Following up on the point that proportions of the population who are obese or
overweight reflect to an extent arbitrary cutoffs, Table 12 presents FGT measures of
obesity using BMI in both years. The FGT for a=1 shows an average obesity gap of
about 13.5 percent of the cutoff rate for BMI (30). This corresponds to about 10 kilos
for the average obese person in the population.
We now turn to a regression analysis of the characteristics associated with weight gain
and weight loss. Table 13 reports two regressions. First, for the obese population in
2002, we study the variables which affect who loses weight by 2005. Second, for the
non obese population in 2002, we study who gains weight and becomes obese by 2005.
Note there are a number of public programs which provides information about the
health problems associated with being obese including IMSS through its PREVENIMSS
program, the Seguro Popular program to those workers in the informal sector and the
Oportunidades program. In addition to socio-economic variables, we include simple
dummies representing affiliation to IMSS, the receipt of Seguro Popular and receipt of
Oportunidades, in a first exploration of how these programs may interact with obesity.
Selection however in who is affiliated or receives public programs is likely to be
endogenous to obesity levels and in fact none of the three dummy variables are
statistically significant. Table 13 shows that men are more likely than women to lose
19
weight over time although they are not less likely to gain weight over time than women.
Few other variables are significant although having a bike reduces the probability of
gaining weight over time whereas having a vehicle increases the probability of gaining
weight, suggestive of a potential role of exercise in obesity in Mexico.
Table 14 presents a preliminary analysis of mental health indicators by the four groups
of individuals. Table 14 shows that those individuals who are obese continuously have
higher proportions of symptoms of mental illness than any of the other three groups, for
instance as demonstrated by feeling less useful, feeling sad, feeling pessimistic and
feeling fear. We also show how mental health conditions change over time for the four
groups. Overall, the mental health symptoms show reductions over time, perhaps due to
trend effects such as economic growth. It is interesting to note however that the group of
individuals who are obese in 2002 but lose weight over time is the group that shows the
largest reductions in a number of symptoms of mental health. Of course this does not
necessarily mean that losing weight led to an improvement in mental health, the reverse
could also be true, for instance that improvements in mental health enabled weight loss
to occur. In any case, it is suggestive of a relationship between obesity and mental
health.
We now turn to an exploratory analysis of the possible causality between weight and
mental health.
We carry out placebo regressions of the impact of mental health
indicators in 2002 on weight gain between 2002 and 2005. Table 15 presents the results
(we present only the coefficients on the mental health indicators) and shows that for
practically all of the 14 different mental health indicators for both men and women,
20
there is no significant relationship of mental health in 2002 to weight changes between
2002 and 2005.
We now estimate individual fixed effects models relating changes in weight to changes
in mental health indicators between 2002 and 2005, disaggregating by gender, and
urban/rural areas. We also carry out a separate analysis for women in urban areas, by
schooling levels and by age to study whether mental health is more affected by weight
gain by different social groups.
We focus on two general specifications, one using weight (continuous) as the relevant
independent variable and the other using whether an individual is obese or not. Graphics
1 and 2 show that mental health is not always continuously increasing or decreasing
with weight and in particular, in some cases individuals with low BMI have higher
indexes of mental health symptoms than obese individuals (for instance, in feeling
useless).
Table 16 shows results from the individual fixed effects estimation for the overall
sample and by gender and area of residence. Little evidence of a negative effect of
weight gain on mental health is observed. In fact, the few significant coefficients are
suggestive that an increase in weight reduces mental health symptoms. When using
obesity as the independent variable, the results suggest no significant relationship
between obesity and mental health.
Table 17 presents the same regressions for women living in urban areas by level of
schooling, dividing the sample for women with no schooling, incomplete elementary
21
school, incomplete secondary and completed secondary and above. Again, the overall
picture is of few significant relationships between obesity and mental health although
there are a couple positive and significant coefficients of weight gain on mental health
are positive, in accordance with increases in obesity reducing mental health. Table 18
presents similar regressions for urban women by age group and again shows few
significant coefficients.
7. Conclusions and Policy Implications.
In this paper, we have studied several facets of health. First, we have used the MxFLS
to provide a broad description of health levels of the Mexican population and how
health indicators are evolving overtime. This analysis confirmed the very high levels of
obesity of the Mexican population particularly that of women, as well as high rates of
high blood pressure and diabetes. On the other hand, however, there are some
noticeable improvements in health conditions over time by the Mexican population in
conditions such as colds, coughs, diarrhea and headaches, suggesting some
improvements in health along these lines and particularly in rural areas.
Secondly, we have used the MxFlS surveys to analyze reporting errors in health,
analyzing for a number of variables including weight, height blood pressure, and
diabetes the extent to which what individuals report correlates with actual measures in
the Mexican context. Reporting errors may derive from lack of knowledge about true
health conditions or embarrassment in reporting, such as in the case of weight. We have
found that with respect to weight, obese individuals tend to substantially underreport
their weight and over-report their height. With respect to diabetes and hypertension, a
large fraction of the population is unaware or at least reports being unaware they have
22
these important chronic diseases. This is clearly an important area for public policy
intervention, diseases such as diabetes and hypertension are among the most serious
health threats among the Mexican adult population and likely to grow in importance
overtime. Public policy institutions should consider more aggressive health campaigns
to encourage individuals to be tested for these diseases and to enter treatment if found to
be ill.
Finally, we have begun to explore in more detail the topic of obesity in Mexico,
presenting data on the persistence of obesity over time and its potential relationship with
mental health indicators. Our descriptive analysis shows that a surprising amount of
transitions in and out of obesity with about 15 percent of the population either entering
obesity or exiting obesity overtime. We also show that few socio-economic variables
are associated with weight loss or weight gain, although men are more likely to lose
weight over time than women. Finally, we carry out some suggestive descriptive
showing that those individuals who lose weight over time tend to have greater
improvements in mental health indicators than other individuals, although the regression
analysis shows more ambiguous results.
The level of obesity in Mexico is among the highest in the world and it is clearly an
urgent public policy intervention is needed. While losing weight is a notably difficult
outcome to achieve, public policy institutions should at least aim to reduce the
probability of children becoming obese and for those adults already overweight, aim for
adults to maintain their body weight, rather than increasing. Further research is needed
to address the best way to avoid obesity in the population, but potential policy options
include information campaigns on the risks associated with obesity, as well as other
programs to, promote exercise and promote healthier eating such as eliminating junk
23
food from schools etc. The results here provide some evidence that at least a small
fraction of the population loses a significant amount of weight over time so at least for a
part of the population, weight loss appears to be possible.
24
References
Behrman, Jere R. & Deolalikar, Anil B., 1988. "Health and nutrition," Handbook of
Development Economics, in: Hollis Chenery & T.N. Srinivasan (ed.), Handbook of
Development Economics, edition 1, volume 1, chapter 14, pages 631-711, Elsevier.
Christakis NA, Fowler JH.The Spread of Obesity in a Large Social Network over 32
Years. New England Journal of Medicine. 2007 Jul 26;357(4):370-9
Cuevas, F., L. Rubalcava and G. Teruel. 2006. Crime Load and Mental Health Decline:
Longitudinal Evidence from the Mexican Family Life Survey. California Center for
Population Research working paper..
Cutler, David M, Edward L. Glaeser, Jesse M. Shapiro Why Have Americans Become
More Obese? The Journal of Economic Perspectives, Vol. 17, No. 3 (Summer, 2003),
pp. 93-118.
Deaton, Angus. Income, aging, health and wellbeing around the world: Evidence from
the Gallup World Poll. Mimeo. 2007.
Felton, Andrew and Carol Graham. 2006. “Variance in Obesity across Cohorts and
Countries: A Norms-Based Explanation using Happiness Surveys.” Mimeo. The
Brookings Institution.
Fernald Lia, Juan Pablo Gutierrez, Lynnette M. Neufeld, Gustavo Olaiz, Stefano M.
Bertozzi, Michele Mietus-Snyder, and Paul J. Gertler. High Prevalence of Obesity
Among the Poor in Mexico. JAMA, June 2, 2004; 291: 2544 - 2545.
Haas JD, Brownlie T. Iron deficiency and reduced work capacity: a critical review of
the research to determine a causal relationship. J Nutr 2001;(supplement) 131:676S688S.
Iburg KM, Salomon JA, Tandon A, Murray CJ. Cross-population comparability of
physician-assessed and self-reported measures of health. In: Murray CJ, Salomon JA,
Mathers C, Lopez AD, editors. Summary Measures of Population Health: Concepts,
Ethics,
King G, Murray CJL, Salomon JA, Tandon A. Enhancing the validity and
cross-cultural comparability of measurement in survey research. Am Polit Sci Rev
2004;98:191-207.
Kyffin, RG, MJ Goldacre, M Gill. Mortality rates and self reported health: database
analysis by English local authority area. BMJ 2004;329:887-888 (16 October).
Manski, Charles F, 1993. "Identification of Endogenous Social Effects: The
Reflection Problem," Review of Economic Studies, Blackwell Publishing, vol. 60(3),
pages 531-42, July.
25
McElroy SL, Kotwal R, Malhotra S, Nelson EB, Keck PE, Nemeroff CB. J Are mood
disorders and obesity related? A review for the mental health professional. Clin
Psychiatry. 2005 Feb;66(2):265.
Murray CJ, Shengelia B, Gupta N, Moussavi S, Tandon A, Thieren M. Validity of
reported vaccination coverage in 45 countries. Lancet 2003 September
27;362(9389):1022-7.
Palloni A, Ewbank DC. Selection process in the study of racial and ethnic differentials
in adult health and mortality. In: Anderson NB, Bulatao RA, Cohen B, editors. Critical
perspectives on racial and ethnic differences in health in late life: The National
Academies Press; 2004. p. 177-226.
Salomon JA. Reconsidering the use of rankings in the valuation of health states: a
model for estimating cardinal values from ordinal data. Popul Health Metr 2003
December 19;1(1):12.
Salomon JA, Tandon A, Murray CJ. Comparability of self rated health: cross sectional
multi-country survey using anchoring vignettes. BMJ 2004 January 31;328(7434):258.
Rahman MO, AJ Barsky. Self-Reported Health Among Older Bangladeshis: How
Good a Health Indicator Is It? The Gerontologist 43:856-863 (2003)
Ren XS, Amick BC. Racial and ethnic disparities in self-assessed health status:
evidence from the National Survey of Families and Households. Ethn Health
1996;1:293-303.
Robinson, F., and J. Keithley (2000), “The Impacts of Crime on Health and Health
Services: A Literature Review”, Health, Risk & Society, vol.2, No.3, 253-266.
Rubalcava, L., and G. Teruel (2006), “User’s Guide for the Mexican
Family Life Survey: The First Wave”, available at http://www.mxfls.cide.edu.
Sadana R, Mathers C, Lopez AD, Murray CJ, Iburg KM. Comparative analyses of more
than 50 household surveys on health status. In: Murray CJ, Salomon JA, Mathers C,
Lopez AD, editors. Summary Measures of Population Health: Concepts, Ethics,
Sadana, R. (2001). "Self report versus observed health status: Evaluation of crosspopulation comparability," Prepared for Global conferences on summary measures of
population health, Geneva: World Health Organization.
Salomon JA, Mathers C, Charette C, Sadana R, Ustun B, Murray CJL. Quantifying
Individual Levels on Health: Definitions, Concepts, and Measurement Issues. In:
Murray CJL, Evans DB, editors. Health Systems Performance Assessment: debates,
methods and empiricism.Geneva: World Health Organization; 2003. p. 301-18.
Secretaria de Salud “Salud en México 2004”, Secretaria de Salud, México DF, 2004.
26
Secretaria de Salud “Salud en México 2001-2005. Información para la rendición de
cuentas”, Secretaria de Salud, México DF, 2005.
Sen, Amartyz. 2002. Health: perception versus observation. Editorial British Medical
Journal.;324:860-861 (13 April ).
Sen A 1994. Objectivity and position: assessment of health and well-being. In Health
and social change ininternational perspective, LC Chen, A Kleinman, NC Ware (eds.).
Harvard series on population and international health, Harvard School of Public Health,
Boston, pp. 115-128.
Strauss J, Thomas D. Health, nutrition, and economic development. J Econ Lit
1998;36:766-817.
Strauss J, Thomas D. Measurement and mismeasurement of social indicators.
Am.Econ.Rev. 1996 May;86:30-34.
Teruel, G., L. Rubalcava, D. Thomas, N. Goldman. (2007). Do Healthier Mexicans
Migrate to the United States? New Findings from the Mexican Family Life Survey.
American Journal of Public Health, forthcoming.
Thomas D and Frankenberg E. 2000. The measurement and interpretation of health in
social surveys. In CJL Murray, CD Mathers, AD Lopez, J Salomon, R Lozano (eds).
Summary measures of population health, Geneva, World Health Organization.
Ware JE, Brook RH, Davies AR, Lohr KN. Choosing measures of health status for
individuals in general populations. Am J Public Health 1981;71:620-625.
Wong, R, M Peláez, and A Palloni. 2005. “Self-reported Global Health among Older
Adults in Latin America and the Caribbean: usefulness of the indicator,” Panamerican
Journal of Public Health 17(5/6):323-332.
World Health Organization. Iron deficiency anemia, assessment, prevention and
control: a guide for program managers. 2001;WHO/NHD/01.3.
27
Table 1. Attrition and Missing Observations. MxFLS 2002 and 2005.
With pregnant women
Total
Between 20 and 60 years old in 2002
With Anthropometric data
With Activities of Daily Living (Over 50 years) data
With Chronic Diseases data
With Hospitalization data
With Medical Assistance data
Without pregnant women in 2002, 2005 or in
between
In 2002
and not in
2005
In 2005 and
not in 2002
In 2002 and
2005
In 2002 and
not in 2005
In 2005 and
not in 2002
In 2002 and 2005
3,339
1,718
1,109
169
1,072
1,356
1,706
4,355
NA
NA
NA
NA
NA
NA
32,370
15,246
12,185
2,394
11,569
13,984
15,106
3,279
1,692
1,083
169
1,048
1,331
1,339
4,278
NA
NA
NA
NA
NA
NA
31,175
14,238
11,244
2,394
10,684
12,993
13,050
28
Table 2a. Longitudinal Information on health: Individuals between 20 and 60 years old in 2002.
2002
2005
Mean
SD
N
Mean
SD
N
Mean
Difference
SD
N
10317
0.77
4.96
8713
-
-
-
1.10
8.40
8728
Physic and anthropometric measures
Height (cm)
159
9.68
11271
Hip measure (cm)
100
10.56
11274
Weight (kg)
70
14.40
11244
Body Mass Index (Weight/(Height*Height))
160
71
9.59
14.71
10355
28
5.15
11157
28
5.23
10287
0.17
3.43
8606
Overweight (30>BMI>25)
40%
0.49
11157
41%
0.49
10287
0.02
0.52
8606
Obese (BMI>30)
28%
0.45
11157
29%
0.46
10287
0.01
0.38
8606
Blood Pressure (systole)
129
22.25
11172
118
15.34
10312
-11
22.26
8630
8583
Blood Pressure (diastole)
80
13.74
11114
77
11.13
10312
-4
15.41
High Blood Pressure (Between 120/80 and 140/90)
20%
0.60
8561
37%
0.48
8587
0.17
0.38
8561
Hypertension (Over 140/90)
16%
0.36
8587
7%
0.25
8587
-0.09
0.40
8587
Hemoglobin measure (Hb/dL, only women between 20-50 in 2005)
12.8
1.92
2110
0.72
2.02
2110
11789
13.5
1.80
2110
High level of glucose (Over 186 mg/dL without 8 hr fasting , under 60 in 2005)
10%
5%
0.31
0.21
4438
2128
High level of cholesterol (Over 240mg/dL, under 60 in 2005)
5%
0.22
5630
High level of glucose (Over 126 mg/dL with 8 hr fasting , under 60 in 2005)
Health perception
Good health (Good or very good)
50%
0.50
13050
55%
0.50
12690
0.05
0.61
Bad health (Bad or very bad)
5%
0.22
13050
4%
0.20
12690
-0.01
0.27
11789
Good health compared with others (Good or very good)
38%
0.49
13051
37%
0.48
12680
-0.01
0.64
11782
Bad health compared with others (Bad or very bad)
Health in previous 4 weeks
6%
0.24
13051
6%
0.24
12680
0.00
0.32
11782
8%
0.28
12963
7%
0.26
12626
-0.01
0.36
11657
8
8.80
1092
8
8.19
898
0.80
11.58
167
20%
0.40
13044
13%
0.34
9944
-0.08
0.49
9906
Stopped work because of sickness
Days not working because of illness
Serious health problems
Medical insurance
IMSS
35%
0.48
11954
34%
0.47
11272
-0.02
0.44
9929
ISSSTE
7%
0.26
11954
7%
0.26
11270
0.00
0.20
9927
Other
3%
0.17
11954
7%
0.26
11272
0.05
0.28
9929
11706
In the past 4 weeks, did you have
Flu
24%
0.61
13048
18%
0.52
12609
-0.05
0.76
Cough
20%
0.59
13049
14%
0.49
12606
-0.05
0.74
11704
Difficulty breathing
12%
0.56
13048
8%
0.44
12607
-0.04
0.68
11704
Stomach ache
20%
0.61
13050
14%
0.50
12605
-0.05
0.75
11704
.ausea
11%
0.55
13050
8%
0.44
12606
-0.03
0.69
11705
11706
Diarrhea
12%
0.59
13050
7%
0.45
12607
-0.04
0.72
Swollen joints
16%
0.60
13050
13%
0.50
12607
-0.03
0.75
11706
Rash
12%
0.58
13050
8%
0.46
12607
-0.03
0.71
11706
Tooth ache
18%
0.62
13050
12%
0.49
12607
-0.06
0.77
11706
Irritated eyes
24%
0.64
13050
17%
0.52
12607
-0.06
0.79
11706
Headache
35%
0.67
13050
28%
0.58
12607
-0.07
0.83
11706
Fever
11%
0.57
13050
9%
0.46
12607
-0.02
0.72
11706
Chest’s pain
12%
0.57
13050
8%
0.47
12607
-0.03
0.71
11706
Body ache
26%
0.65
13050
21%
0.55
12607
-0.05
0.82
11706
Go to the bathroom frequently
25%
0.79
13050
19%
0.66
12692
-0.06
0.95
11791
Carry heavy bucket
65%
0.52
2394
61%
0.49
3347
-0.08
0.62
2195
Walk 5 Kilometers
62%
0.53
2394
59%
0.53
3346
-0.07
0.65
2195
Kneel down
64%
0.53
2394
63%
0.48
3347
-0.05
0.62
2195
Claim stairs
77%
0.50
2394
77%
0.46
3347
-0.04
0.60
2195
Able to get dressed without help
95%
0.27
2394
92%
0.27
3347
-0.04
0.38
2195
Stand on a chair
90%
0.36
2394
85%
0.36
3347
-0.08
0.49
2195
Go to the toilet without help
96%
0.25
2394
93%
0.26
3347
-0.05
0.36
2195
Stand up
80%
0.46
2394
75%
0.43
3347
-0.08
0.58
2195
Health status, only for older than 50 years. Are you capable of
Chronic Disease. Do you have
Diabetes
6%
0.24
10684
7%
0.25
11261
0.01
0.21
8897
Arterial hypertension
11%
0.31
10684
9%
0.29
11261
-0.02
0.34
8897
Heart disease
2%
0.15
10684
2%
0.13
11260
-0.01
0.17
8896
Cancer
1%
0.08
10684
1%
0.08
11261
0.00
0.10
8897
Arthritis o Rheumatism
4%
0.21
10684
3%
0.17
11261
-0.02
0.24
8897
Gastric Ulcer
8%
0.27
10684
5%
0.21
11261
-0.03
0.31
8897
Migraine
4%
0.19
10684
2%
0.15
11261
-0.01
0.21
8897
Medical assistance during the last 4 weeks
16%
0.37
13050
11%
0.31
12623
-0.05
0.46
11721
Number of visits to the doctor
Hospitalization during the last 12 months
0.17
5%
0.45
0.26
13050
12993
0.12
4%
0.40
0.22
12623
12549
-0.05
-0.01
0.56
0.32
11721
11618
Medical assistance
.ote: All the anthropometric indicators are based on the definitions of the .ational Institute of Health. Excludes pregnant women in 2002, 2005 or in between.
29
Table 2b. Change in medical consultations. By type of clinic.
Number of visits to the doctor
Absolute Percentile
2002
2005
Change
Change
SSA
414
329
-85
-21%
IMSS
706
458
-248
-35%
ISSSTE
147
95
-52
-35%
PEMEX
20
27
7
35%
Private-clinic
185
133
-52
-28%
Private-doctor
387
248
-139
-36%
DIF
11
12
1
9%
Nurse
10
4
-6
-60%
Mobil Unit
1
0
-1
-100%
Red Cross
155
13
-142
-92%
Drug deposit
9
13
4
44%
Drugstore
17
57
40
235%
28
26
-2
-7%
Traditional medicine
Other
18
4
-14
-78%
Total
2108
1419
-689
-33%
Observations
11721
11721
% Seeking medical
care
0.18
0.12
-0.059
.ote: Excludes pregnant women in 2002, 2005 or in between.
30
Table 3. 2002 Information on health: Individuals between 20 and 60 years old in 2002.
Rural
Women
Mean
SD
Urban
N
Mean
Men
SD
N
Women
Mean
SD
N
Mean
Men
SD
N
Physic and anthropometric measures
Height (cm)
Hip measure (cm)
Weight (kg)
Body Mass Index (Weight/(Height*Height))
Overweight (30>BMI>25)
Obese (BMI>30)
Blood Pressure (systole)
Blood Pressure (diastole)
High Blood Pressure (Between 120/80 and 140/90)
Hypertension (Over 140/90)
152
101
66
28
34%
35%
126
79
17%
13%
13
7.00
11.92
14.24
5.71
0.47
0.48
22.08
13.12
0.38
0.33
2.00
2471
2476
2474
2446
2446
2446
2446
2434
2094
2102
860
165
96
72
27
42%
19%
133
83
19%
21%
7.63
8.80
14.10
4.50
0.49
0.39
21.17
13.37
0.39
0.41
2220
2229
2218
2199
2199
2199
2205
2192
1638
1643
154
102
67
28
37%
33%
124
77
15%
11%
13
6.99
11.40
13.62
5.60
0.48
0.47
22.31
13.44
0.36
0.31
1.87
3561
3562
3551
3528
3528
3528
3529
3534
2821
2826
1250
167
98
76
27
45%
24%
133
83
18%
21%
7.44
8.45
13.50
4.28
0.50
0.43
21.65
13.97
0.39
0.40
3019
3007
3001
2984
2984
2984
2992
2954
2008
2016
39%
8%
29%
9%
0.49
0.27
0.45
0.29
2751
2751
2752
2752
49%
5%
38%
5%
0.50
0.23
0.49
0.22
2751
2751
2751
2751
50%
5%
39%
8%
0.50
0.22
0.49
0.26
3878
3878
3878
3878
60%
3%
45%
3%
0.49
0.16
0.50
0.18
3670
3670
3670
3670
11%
8
22%
0.32
8.21
0.41
2729
307
2751
6%
10
17%
0.24
9.35
0.38
2726
174
2747
11%
8
24%
0.31
7.96
0.43
3855
413
3878
5%
10
17%
0.23 3653
10.36 198
0.38 3668
22%
5%
1%
0.41
0.22
0.11
2659
2659
2659
24%
4%
1%
0.42
0.21
0.11
2391
2391
2391
44%
10%
4%
0.50
0.30
0.20
3730
3730
3730
46%
8%
4%
0.50
0.27
0.20
3174
3174
3174
28%
24%
15%
24%
14%
13%
19%
13%
21%
23%
46%
14%
15%
32%
25%
0.57
0.55
0.52
0.57
0.51
0.52
0.54
0.50
0.59
0.57
0.64
0.55
0.54
0.60
0.67
2751
2752
2751
2752
2752
2752
2752
2752
2752
2752
2752
2752
2752
2752
2752
29%
25%
16%
23%
13%
17%
17%
18%
23%
30%
32%
17%
17%
28%
32%
0.85
0.82
0.81
0.86
0.82
0.87
0.87
0.87
0.89
0.92
0.91
0.87
0.88
0.90
1.07
2751
2751
2751
2751
2751
2751
2751
2751
2751
2751
2751
2751
2751
2751
2751
22%
18%
12%
19%
12%
9%
19%
10%
16%
21%
42%
9%
10%
28%
24%
0.45
0.42
0.39
0.44
0.40
0.35
0.45
0.39
0.40
0.46
0.54
0.35
0.37
0.49
0.67
3876
3877
3877
3877
3877
3877
3877
3877
3877
3877
3877
3877
3877
3877
3877
19%
16%
9%
14%
7%
11%
11%
9%
15%
23%
23%
7%
8%
19%
21%
0.56
0.57
0.50
0.54
0.47
0.57
0.53
0.52
0.57
0.59
0.57
0.50
0.47
0.58
0.76
3670
3669
3669
3670
3670
3670
3670
3670
3670
3670
3670
3670
3670
3670
3670
Carry heavy bucket
Walk 5 Kilometers
Kneel down
Claim stairs
Able to get dressed without help
Stand on a chair
Go to the toilet without help
Stand up
46%
46%
47%
59%
92%
85%
94%
67%
0.50
0.50
0.50
0.49
0.26
0.36
0.24
0.47
567
567
567
567
567
567
567
567
83%
74%
77%
86%
97%
95%
98%
91%
0.50
0.54
0.53
0.57
0.36
0.39
0.35
0.43
544
544
544
544
544
544
544
544
51%
51%
54%
73%
94%
86%
94%
70%
0.50
0.50
0.50
0.45
0.24
0.34
0.23
0.46
691
691
691
691
691
691
691
691
83%
81%
79%
91%
97%
97%
97%
93%
0.49
0.50
0.51
0.42
0.18
0.36
0.16
0.40
592
592
592
592
592
592
592
592
Diabetes
Arterial hypertension
Heart disease
Cancer
Arthritis o Rheumatism
Gastric Ulcer
Migraine
7%
15%
3%
1%
5%
8%
4%
0.25
0.35
0.17
0.10
0.22
0.26
0.19
2535
2535
2535
2535
2535
2535
2535
4%
5%
2%
0%
4%
6%
1%
0.19
0.22
0.13
0.04
0.20
0.23
0.12
2262
2262
2262
2262
2262
2262
2262
7%
16%
3%
1%
6%
10%
7%
0.26
0.36
0.16
0.11
0.23
0.30
0.25
3190
3190
3190
3190
3190
3190
3190
5%
7%
2%
0%
3%
7%
2%
0.22
0.25
0.13
0.05
0.16
0.26
0.13
2697
2697
2697
2697
2697
2697
2697
Medical assistance
20% 0.40 2752 10% 0.30 2751
21% 0.41 3877 11% 0.32
Medical assistance during the last 4 weeks
0.22
0.53 2752 0.11
0.33 2751
0.22
0.45 3877 0.13
0.44
Number of visits to the doctor
6%
0.25 2740 2%
0.16 2733
8%
0.35 3866 3%
0.20
Hospitalization during the last 12 months
.ote: All the anthropometric indicators are based on the definitions of the .ational Institute of Health. Excludes pregnant women in 2002, 2005 or in between.
3670
3670
3654
Hemoglobin measure (Hb/dL, only women between 20-50 in 2005)
Health perception
Good health (Good or very good)
Bad health (Bad or very bad)
Good health compared with others (Good or very good)
Bad health compared with others (Bad or very bad)
Health in previous 4 weeks
Stopped work because of sickness
Days not working because of illness
Serious health problems
Medical insurance
IMSS
ISSSTE
Other
In the past 4 weeks, did you have
Flu
Cough
Difficulty breathing
Stomach ache
.ausea
Diarrhea
Swollen joints
Rash
Tooth ache
Irritated eyes
Headache
Fever
Chest’s pain
Body ache
Go to the bathroom frequently
Health status, only for older than 50 years Are you capable of
Chronic Disease. Do you have
31
Table 4. Differences 2005 minus 2002. Individuals between 20 and 60 years old in 2002.
Rural
Women
Urban
Men
Women
Men
Mean
SD
N
Mean
SD
N
Mean
SD
N
Mean
SD
N
Height (cm)
0.95
4.85
2141
0.55
5.60
1663
0.89
4.53
2839
0.59
5.06
2070
Weight (kg)
1.13
7.71
2149
1.44
9.52
1675
0.87
8.02
2839
1.10
8.61
2065
Body Mass Index (Weight/(Height*Height))
0.12
3.51
2120
0.35
3.48
1644
0.06
3.53
2801
0.21
3.15
2041
Overweight (30>BMI>25)
0.02
0.50
2120
0.02
0.54
1644
0.00
0.52
2801
0.03
0.52
2041
Obese (BMI>30)
0.00
0.38
2120
0.02
0.39
1644
0.01
0.39
2801
0.00
0.38
2041
Blood Pressure (systole)
-9.44
22.11
2113
-14.0
22.39
1653
-8.06
21.84
2824
-13.0
22.33
2040
Blood Pressure (diastole)
-3.85
14.52
2101
-6.08
15.31
1647
-2.02
15.54
2822
-4.79
15.89
2013
High Blood Pressure (Between 120/80 and 140/90)
0.15
0.60
2094
0.23
0.63
1638
0.18
0.58
2821
0.25
0.62
2008
Hypertension (Over 140/90)
-0.06
0.38
2102
-0.14
0.45
1643
-0.05
0.35
2826
-0.13
0.45
2016
Hemoglobin measure (Hb/dL, women between 20-50 in 2005)
0.69
2.08
860
0.73
1.98
1250
Physic and anthropometric measures
Health perception
Good health (Good or very good)
0.06
0.60
2572
0.08
0.64
2506
0.03
0.60
3473
0.04
0.60
3238
Bad health (Bad or very bad)
Good health compared with others (Good or very good)
-0.02
-0.01
0.33
0.61
2572
2574
-0.01
-0.01
0.27
0.65
2506
2506
-0.01
-0.02
0.27
0.64
3473
3467
0.00
-0.02
0.21
0.66
3238
3235
Bad health compared with others (Bad or very bad)
0.00
0.37
2574
0.00
0.29
2506
0.00
0.35
3467
0.01
0.25
3235
-0.01
0.41
2536
0.00
0.33
2478
-0.03
0.39
3433
-0.01
0.28
3210
1.62
9.78
60
-2.95
12.78
22
1.98
12.27
58
-0.52
12.60
27
-0.07
0.49
2395
-0.09
0.45
1902
-0.08
0.52
3195
-0.08
0.46
2414
Health in previous 4 weeks
Stopped work because of sickness
Days not working because of illness
Serious health problems
Medical insurance
IMSS
0.00
0
2394
-0.01
0.41
1910
-0.02
0
3217
-0.03
0.48
2408
ISSSTE
0.00
0
2394
-0.01
0.16
1910
0.00
0
3217
0.00
0.22
2406
Other
0.09
0
2394
0.07
0.29
1910
0.02
0
3217
0.01
0.24
2408
In the past 4 weeks, did you have
Flu
-0.06
0.72
2556
-0.09
0.96
2492
-0.06
0.57
3452
-0.01
0.81
3206
Cough
-0.08
0.69
2556
-0.08
0.92
2492
-0.05
0.54
3451
-0.02
0.80
3205
Difficulty breathing
-0.05
0.62
2555
-0.06
0.89
2492
-0.03
0.48
3452
-0.01
0.74
3205
Stomach ache
-0.06
0.69
2556
-0.09
0.91
2492
-0.04
0.57
3451
-0.02
0.83
3205
.ausea
-0.04
0.61
2556
-0.04
0.89
2492
-0.03
0.51
3451
0.00
0.73
3206
Diarrhea
-0.04
0.62
2556
-0.07
0.96
2492
-0.03
0.46
3452
-0.03
0.79
3206
Swollen joints
-0.05
0.66
2556
-0.05
0.95
2492
-0.02
0.57
3452
0.00
0.80
3206
Rash
-0.05
0.60
2556
-0.08
0.94
2492
-0.01
0.50
3452
-0.01
0.78
3206
Tooth ache
-0.07
0.70
2556
-0.08
1.02
2492
-0.04
0.50
3452
-0.04
0.81
3206
Irritated eyes
-0.06
0.69
2556
-0.11
1.04
2492
-0.06
0.57
3452
-0.05
0.85
3206
Headache
-0.11
0.77
2556
-0.08
1.03
2492
-0.09
0.69
3452
-0.02
0.84
3206
Fever
-0.03
0.66
2556
-0.05
0.97
2492
0.00
0.45
3452
0.01
0.78
3206
Chest’s pain
-0.05
0.63
2556
-0.06
0.99
2492
-0.02
0.47
3452
0.00
0.73
3206
Body ache
-0.07
0.72
2556
-0.07
1.03
2492
-0.04
0.63
3452
-0.01
0.87
3206
Go to the bathroom frequently
-0.08
0.75
2574
-0.07
1.27
2506
-0.08
0.76
3473
-0.03
1.00
3238
Carry heavy bucket
-8%
0.64
533
-12%
0.62
511
-8%
0.64
628
-3%
0.58
523
Walk 5 Kilometers
-7%
0.61
533
-8%
0.64
511
-7%
0.65
628
-6%
0.68
523
523
Health status, only for older than 50 years. Are you capable of
Kneel down
-4%
0.62
533
-11%
0.64
511
-3%
0.62
628
-1%
0.59
Claim stairs
-1%
0.60
533
-12%
0.67
511
-1%
0.62
628
-3%
0.50
523
Able to get dressed without help
-6%
0.42
533
-7%
0.45
511
-3%
0.37
628
-1%
0.25
523
Stand on a chair
-9%
0.52
533
-9%
0.50
511
-8%
0.49
628
-6%
0.44
523
Go to the toilet without help
-7%
0.39
533
-6%
0.43
511
-3%
0.35
628
-2%
0.24
523
Stand up
-6%
0.63
533
-13%
0.56
511
-6%
0.60
628
-9%
0.52
523
2032
Chronic Disease. Do you have
Diabetes
1%
0.22
2285
1%
0.19
1822
1%
0.23
2758
1%
0.21
Arterial hypertension
-3%
0.40
2285
-1%
0.25
1822
-3%
0.39
2758
-1%
0.28
2032
Heart disease
-1%
0.17
2285
-1%
0.16
1821
-1%
0.19
2758
0%
0.15
2032
Cancer
0%
0.11
2285
0%
0.06
1822
0%
0.14
2758
0%
0.05
2032
Arthritis o Rheumatism
-1%
0.25
2285
-2%
0.24
1822
-2%
0.26
2758
-1%
0.19
2032
Gastric Ulcer
-3%
0.30
2285
-3%
0.28
1822
-4%
0.35
2758
-3%
0.28
2032
Migraine
-2%
0.21
2285
0%
0.13
1822
-2%
0.27
2758
-1%
0.15
2032
Medical assistance
Medical assistance during the last 4 weeks
-0.06
0.51
2557
-0.03
0.38
2495
-0.06
0.52
3455
-0.04
0.40
3214
Number of visits to the doctor
-0.07
0.55
2557
-0.02
0.59
2495
-0.06
0.58
3455
-0.05
0.54
3214
Hospitalization during the last 12 months
-0.01
0.33
2535
-0.01
0.20
2469
-0.02
0.43
3432
-0.01
0.24
3182
.ote: All the anthropometric indicators are based on the definitions of the .ational Institute of Health. Excludes pregnant women in 2002, 2005 or in between.
32
Table 5. Distribution of individuals according to the error reporting weight and height. MxFLS 2002.
Weight
Total
Urban
Rural
Total
Men
Women
Total
Men
Women
Total
Men
Women
% Accurate report
% Under-reporting
66%
19%
60%
20%
71%
18%
66%
18%
60%
20%
72%
17%
65%
19%
59%
19%
70%
19%
% Over-reporting
16%
21%
11%
15%
20%
11%
16%
22%
11%
Height
Total
Urban
Rural
Total
Men
Women
Total
Men
Women
Total
Men
Women
% Accurate report
% Under-reporting
55%
11%
54%
11%
56%
10%
57%
9%
56%
10%
58%
8%
50%
15%
49%
15%
52%
15%
% Over-reporting
34%
35%
34%
34%
34%
34%
35%
36%
33%
Accurate reporting is when the error reporting weight or height is smaller to half SD.
pregnant women in 2002, 2005 or in between.
33
Excludes
Table 6. Distribution of Individuals according to health problems.
Hypertension (MxFLS 2002). Higher than 140/90.
Total
% Correctly identify no health
problem
% With health problem and aware
Urban
Rural
Total
Men
Women
Total
Men
Women
Total
Men
Women
77%
76%
79%
77%
75%
79%
77%
75%
79%
3%
2%
3%
3%
3%
3%
3%
3%
3%
Taking medicine
% No health problem, but believe they
have
Taking medicine
1%
0%
1%
1%
1%
1%
1%
1%
1%
7%
3%
10%
8%
4%
10%
8%
4%
10%
2%
0%
2%
2%
1%
3%
2%
1%
3%
% With health problem but unaware
13%
19%
9%
13%
19%
8%
13%
19%
8%
Diabetes (MxFLS 2005) Definition1. Over 8 hr. fasting, diabetes over 126 (mg/dL) (68%).
Total
Total
% Correctly identify no health
problem
Men
Urban
Women
Total
Men
Rural
Women
Total
Men
Women
86%
86%
86%
85%
85%
86%
86%
87%
86%
% With health problem and aware
4%
3%
4%
4%
4%
4%
3%
2%
4%
Taking medicine
% No health problem, but believe they
have
2%
2%
2%
2%
2%
2%
2%
2%
3%
4%
3%
4%
4%
3%
5%
4%
4%
4%
Taking medicine
1%
1%
1%
1%
2%
1%
1%
1%
1%
7%
8%
6%
6%
8%
5%
7%
7%
6%
% With health problem but unaware
Diabetes (MxFLS 2005) Definition2. Less 8 hr. fasting, diabetes over 186 (mg/dL) (32%).
Total
Total
% Correctly identify no health
problem
Men
Urban
Women
Total
Men
Rural
Women
Total
Men
Women
90%
93%
89%
90%
91%
88%
91%
94%
89%
% With health problem and aware
2%
2%
3%
3%
3%
3%
2%
1%
2%
Taking medicine
% No health problem, but believe they
have
1%
0%
2%
1%
0%
2%
1%
0%
1%
5%
3%
6%
6%
4%
7%
4%
2%
6%
Taking medicine
2%
1%
2%
1%
1%
2%
2%
0%
3%
2%
2%
3%
2%
2%
2%
2%
2%
3%
% With health problem but unaware
Hypertension and Diabetes are defined according to the .ational Institute of Health. Excludes pregnant women in 2002, 2005 or
in between.
34
Table 7. Error reporting Weight (Reported minus Measured). MxFLS 2002.
OLS
Probit
OLS
Probit
1
2
1
2
Error
reporting
Weight
Accurate
report (Error
smaller than
half SD)
Error
reporting
Weight
Accurate
report (Error
smaller than
half SD)
-0.159
-0.026
-0.155
0.029
[0.221]
[0.020]
[0.181]
[0.017]
-0.124
0.006
[0.246]
[0.022]
0.477
-0.006
[0.274]
[0.025]
0.462
-0.118
[0.205]*
[0.019]**
0.022
0.005
[0.029]
[0.003]
Individual Characteristics
Household Services
Between 30 and 39
Between 40 and 49
Between 50 and 59
Gender (Men=1)
Cognitive Capacity
Speak indigenous language
-1.074
-0.023
[0.333]**
[0.031]
Phone
Electricity
House were they live
Other houses
Electronic devices
Bicycle
0.704
-0.05
[0.400]
[0.038]
0.555
-0.016
[0.287]
[0.027]
0.676
0.009
[0.214]**
[0.020]
Divorce
Between 15 and 100 thousand
habitants
-0.009
0.036
Separate
[0.271]
[0.024]
Between 2.5 and 15 thousand
habitants
-0.076
-0.024
[0.271]
[0.025]
-0.115
-0.002
[0.204]
[0.019]
-0.077
0.009
[0.081]
[0.008]
0.013
-0.001
[0.058]
[0.005]
-0.348
-0.001
[0.250]
[0.023]
0.364
-0.052
[0.265]
[0.024]*
Incomplete Elementary School
Incomplete Junior High School
Vehicle
Number of rooms for sleeping
Number of children
A member is a Oportunidades
beneficiary
Domestic Partnership
Married
Single
0.217
-0.01
[0.213]
[0.020]
-0.025
-0.003
[0.183]
[0.017]
0.175
0.008
[0.411]
[0.038]
0.268
0.022
[0.157]
[0.014]
0.263
0.014
[0.174]
[0.016]
0.263
-0.018
[0.583]
[0.054]
1.029
-0.072
[0.947]
[0.091]
0.841
0.025
[0.681]
[0.061]
0.292
0.034
[0.541]
[0.050]
0.245
0.035
[0.580]
[0.052]
Body Mass Index
Obese (BMI>30)
Overweight
(30>BMI>25)
Constant
-2.469
-0.061
[0.210]**
[0.020]**
-1.498
0.003
[0.189]**
[0.017]
0.472
[1.074]
Health indicators
Hospitalization previous 12
months
[0.077]
Marital Status
Size of locality
Less than 2.5 thousands
habitants
0.092
[0.801]
Assets
Grades of schooling
.one
-0.897
Observations
4762
4762
0.284
0.094
R-squared
0.04
Seeing a doctor in previous 4
weeks
[0.194]
[0.017]**
Half of the SD of the difference between reported and measured weight was used to define the size of the error in col. 2. SE in
brackets * significant at 5%; ** significant at 1%. Excludes pregnant women in 2002, 2005 or in between. Excluded categories:
more than junior high, more than 100 thousand habitants, widow, and BMI lower than 25.
35
Table 8. Error reporting Height (Reported minus Measured). MxFLS 2002.
OLS
1
Error
reporting
Height
Probit
OLS
2
Accurate
report (Error
smaller than
half SD)
1
Error
reporting
Height
Individual
Characteristics
Probit
2
Accurate
report (Error
smaller than
half SD)
Household Services
Between 30 and 39
Between 40 and 49
Between 50 and 59
Gender (Men=1)
Cognitive Capacity
0.057
0
[0.256]
[0.024]
0.173
0.022
[0.289]
[0.027]
0.437
-0.053
[0.336]
[0.031]
-0.671
-0.058
[0.256]**
[0.024]*
-0.045
0.012
[0.034]
[0.003]**
Phone
Electricity
.one
Incomplete Elementary
School
Incomplete Junior High
School
-1.333
-0.07
[0.452]**
[0.042]
House were they live
Other houses
Electronic devices
Bicycle
0.205
-0.056
[0.592]
[0.055]
Vehicle
-0.575
-0.003
Between 2.5 and 15
thousand habitants
Less than 2.5 thousands
habitants
[0.350]
[0.032]
-0.28
-0.036
[0.227]
[0.021]
Divorce
-0.018
-0.091
Separate
[0.309]
[0.028]**
0.861
-0.069
[0.324]**
[0.030]*
-0.25
-0.037
[0.246]
[0.023]
Number of rooms for
sleeping
-0.153
-0.002
[0.095]
[0.009]
Number of children
-0.138
-0.019
[0.086]
[0.008]*
-0.029
-0.068
[0.344]
[0.032]*
A member is a
Oportunidades
beneficiary
-0.749
0.204
[0.974]
[0.086]*
-0.127
-0.002
[0.248]
[0.023]
-0.087
0.039
[0.221]
[0.020]
-0.651
0.007
[0.560]
[0.052]
0.229
0.009
[0.187]
[0.017]
0.115
0.016
[0.203]
[0.019]
Marital Status
Domestic Partnership
Size of locality
Between 15 and 100
thousand habitants
0.022
[0.019]
Assets
Speak indigenous
language
Grades of schooling
-0.246
[0.210]
Married
Single
-0.947
0.129
[0.842]
[0.073]
-0.739
0.057
[1.127]
[0.102]
-0.897
0.094
[0.935]
[0.082]
-0.854
0.149
[0.796]
[0.074]*
-1
0.105
[0.830]
[0.075]
Body Mass Index
Obese (BMI>30)
Overweight
(30>BMI>25)
Constant
1.284
-0.085
[0.253]**
[0.023]**
0.639
-0.037
[0.220]**
[0.020]
4.572
[1.396]**
Health indicators
Hospitalization previous
12 months
-0.008
0.015
[0.328]
[0.031]
Seeing a doctor in
previous 4 weeks
-0.511
0.045
[0.241]*
[0.022]*
Observations
3660
R-squared
0.02
3660
Half of the SD of the difference between reported and measured height was used to define the size of the error in col. 2. SE
in brackets * significant at 5%; ** significant at 1%. Excludes pregnant women in 2002, 2005 or in between. Excluded
categories: more than junior high, more than 100 thousand habitants, widow, and BMI lower than 25.
36
Table 9. Under-reporting hypertension. MxFLS 2002.
1
2
1
Under reporting. Not aware of
hypertension
Without
controlled
With
controlled
-0.062
-0.002
[0.042]
[0.042]
-0.107
-0.043
[0.043]*
[0.043]
Individual Characteristics
2
Under reporting. Not aware of
hypertension
Without
controlled
With
controlled
-0.043
-0.056
[0.026]
[0.030]
0.122
-0.01
[0.108]
[0.116]
-0.035
-0.033
[0.027]
[0.036]
0.029
0.032
[0.023]
[0.029]
0.021
0.034
[0.054]
[0.062]
Household Services
Between 30 and 39
Between 40 and 49
Between 50 and 59
Gender (Men=1)
-0.188
-0.155
[0.051]**
[0.046]**
0.17
0.336
[0.029]**
[0.030]**
0.026
0.107
[0.034]
[0.042]*
Phone
Electricity
Assets
House were they live
Other houses
Speak indigenous language
Grades of schooling
.one
Incomplete Elementary School
Incomplete Junior High School
Electronic devices
0.042
0.062
[0.039]
[0.053]
Bicycle
-0.005
0.033
[0.039]
[0.047]
Vehicle
-0.021
-0.031
Marital Status
[0.034]
[0.041]
Domestic Partnership
-0.042
0.062
[0.039]
[0.041]
-0.026
0.036
[0.037]
[0.041]
Size of locality
Between 15 and 100 thousand
habitants
Between 2.5 and 15 thousand
habitants
Less than 2.5 thousands habitants
Number of rooms for sleeping
Number of children
A member is a Oportunidades
beneficiary
Divorce
Separate
-0.011
0.026
[0.028]
[0.033]
Married
0.007
0.006
[0.011]
[0.013]
Single
0
-0.002
[0.006]
[0.008]
0.004
-0.021
[0.030]
[0.038]
Seeing a doctor in previous 4
weeks
-0.033
-0.081
[0.032]
[0.042]
-0.228
-0.244
[0.036]**
[0.034]**
-0.007
[0.026]
-0.012
-0.023
[0.023]
[0.028]
-0.099
-0.004
[0.086]
[0.076]
0.027
0.029
[0.112]
[0.133]
-0.03
0.066
[0.087]
[0.082]
-0.04
0.019
[0.059]
[0.070]
0.028
0.135
[0.063]
[0.066]*
-0.05
0.01
[0.033]
[0.037]
Body Mass Index
Obese (BMI>30)
Overweight
(30>BMI>25)
Health indicators
Hospitalization previous 12
months
0.009
[0.021]
Observations
-0.004
0.035
[0.032]
[0.036]
1357
1720
Hypertension is defined according to the .ational Institute of Health (higher than 140/90).Controlled are individuals that report
having hypertension but their measured blood pressure did not. SE in brackets * significant at 5%; ** significant at 1%.
Excludes pregnant women in 2002, 2005 or in between. Excluded categories: more than junior high, more than 100 thousand
habitants, widow, and BMI lower than 25.
37
Table 10. Error reporting Diabetes. MxFLS 2005.
Under reporting. Not aware of diabetes
Over 126 mg/dL with 8 hr fasting
Over 186 mg/dL without 8 hr fasting
Without controlled
With controlled
Without controlled
With controlled
Individual Characteristics
Between 30 and 39
Between 40 and 49
Between 50 and 59
Gender (Men=1)
Speak indigenous language
-0.238
[0.198]
-0.512
[0.146]**
-0.53
[0.158]**
0.215
[0.066]**
0.02
[0.118]
-0.278
[0.115]*
-0.52
[0.109]**
-0.478
[0.097]**
0.187
[0.061]**
-0.113
[0.088]
0.978
[0.042]**
1
[0.000]**
0.99
[0.041]**
-0.493
[0.282]
0.569
[0.307]
0.072
[0.302]
-0.084
[0.263]
-0.241
[0.194]
-0.064
[0.089]
-0.099
[0.117]
0.007
[0.131]
-0.071
[0.119]
-0.04
[0.110]
0.006
[0.113]
0
[0.100]
-0.016
[0.092]
0.77
[0.276]**
0.609
[1.062]
0.868
[0.515]
0.228
[0.216]
0.164
[0.175]
0.091
[0.152]
-0.282
[0.097]**
-0.101
[0.098]
0.034
[0.068]
-0.013
[0.024]
0.037
[0.018]*
-0.052
[0.079]
-0.217
[0.071]**
-0.033
[0.082]
0.039
[0.060]
0.011
[0.022]
0.024
[0.016]
0.001
[0.066]
0.789
[0.128]**
0.512
[0.293]
0.053
[0.335]
-0.069
[0.181]
0.127
[0.106]
0.216
[0.300]
0.179
[0.181]
0.179
[0.164]
0.186
[0.107]
0.06
[0.043]
-0.019
[0.021]
0.05
[0.118]
-0.187
[0.090]*
-0.315
[0.079]**
-0.194
[0.089]*
-0.334
[0.048]**
-0.498
[0.140]**
-0.253
[0.054]**
0.016
[0.030]
0.299
[0.299]
0.009
[0.026]
0.034
[0.297]
0.396
[0.225]
-0.012
[0.040]
-0.023
[0.082]
0.051
[0.064]
-0.012
[0.124]
0.017
[0.055]
0.027
[0.060]
-0.061
[0.077]
0.052
[0.060]
-0.034
[0.111]
0.029
[0.048]
0.024
[0.053]
-0.502
[0.269]
0.842
[0.133]**
0.075
[0.359]
0.367
[0.348]
0.46
[0.295]
-0.357
[0.138]**
0.068
[0.112]
-0.006
[0.127]
-0.02
[0.076]
0
[0.086]
0.049
[0.141]
0.064
[0.195]
0.218
[0.112]
0.022
[0.125]
0.165
[0.121]
-0.046
[0.137]
0.043
[0.202]
-0.091
[0.145]
-0.047
[0.120]
0.106
[0.151]
-0.725
[0.107]**
-0.179
[0.093]
-0.562
[0.119]**
-0.982
[0.019]**
-0.174
[0.095]
-0.252
[0.212]
-0.098
[0.154]
0.194
[0.071]**
0.133
[0.073]
407
0.224
[0.064]**
0.156
[0.068]*
566
0.181
[0.475]
-0.642
[0.225]**
71
-0.128
[0.105]
-0.037
[0.107]
166
Grades of schooling
.one
Incomplete Elementary School
Incomplete Junior High School
Size of locality
Between 15 and 100 thousand
habitants
Between 2.5 and 15 thousand
habitants
Less than 2.5 thousands habitants
Number of rooms for sleeping
Number of children
A member is a Oportunidades
beneficiary
Health indicators
Hospitalization previous 12 months
Seeing a doctor in previous 4 weeks
Household Services
Phone
Electricity
Assets
House were they live
Other houses
Electronic devices
Bicycle
Vehicle
Marital Status
Domestic Partnership
Divorce
Separate
Married
Single
Body Mass Index
Obese (BMI>30)
Overweight (30>BMI>25)
Observations
All observations under 60 in 2005. Diabetes is defined according to the .ational Institute of Health. Controlled are individuals
that report having hypertension but their measured blood pressure did not. SE in brackets * significant at 5%; ** significant at
1%. Excludes pregnant women in 2002, 2005 or in between. Excluded categories: more than junior high, more than 100
thousand habitants, widow, and BMI lower than 25.
38
Table 11. Distribution of Individuals according to obesity.
Obesity dynamics (MxFLS 2002 & 2005)
Total
Urban
Rural
Total Men Women Total Men Women Total Men Women
% Obese in 2002 & 2005
% Obese in 2002 & not in 2005
22% 15%
7% 7%
28%
7%
22% 16%
7% 7%
27%
7%
22% 14%
7% 7%
29%
7%
% Not obese in 2002 & obese in 2005
% Not obese in 2002 & 2005
8% 8%
63% 70%
8%
57%
8% 7%
63% 69%
8%
58%
8% 8%
63% 71%
7%
56%
.ote: Excludes pregnant women in 2002, 2005 or in between.
39
Table 12. FGT measures of obesity (BMI>30). By gender, level of education and size of
community.
FGT α=1
FGT α=2
FGT α=0
2002 2005 Change 2002 2005 Change
Total
0.280
0.135 0.139 0.004 0.034 0.037 0.003
Female
Male
Level of education
0.326
0.219
0.153 0.154
0.103 0.109
0.001
0.007
0.043 0.044
0.019 0.024
0.001
0.005
0.317
0.327
0.314
0.224
0.153
0.146
0.130
0.123
-0.011
-0.005
0.004
0.015
0.044
0.041
0.031
0.028
0.038
0.038
0.035
0.038
-0.006
-0.003
0.003
0.010
0.273
0.137 0.138 0.000 0.036 0.036
Female
0.337
0.150 0.150 0.000 0.042 0.041
Male
0.190
0.112 0.113 0.001 0.024 0.027
0.286
0.133 0.140 0.007 0.033 0.038
Female
0.319
0.154 0.157 0.003 0.043 0.047
0.241
0.097 0.107 0.010 0.017 0.022
Male
.ote: The FGT a=1 is calculated with the absolute value (all are negative).
0.000
-0.002
0.004
0.005
0.004
0.005
.one
Incomplete Elementary School
Incomplete Junior High School
High School or more
Rural
Urban
40
0.142
0.141
0.135
0.138
41
Table 13. Relationship between individual characteristics and change in obesity.
1
2
1
Stop
Stop being Start being
being
obese
obese
obese
Individual Characteristics
Health indicators
-0.066
0.016
-0.019
Hospitalization previous
Between 30 and 39
[0.027]*
[0.012]
12 months
[0.033]
0.008
0.019
-0.018
Seeing a doctor in
Between 40 and 49
[0.030]
[0.014]
previous 4 weeks
[0.022]
0.015
0.016
Household Services
Between 50 and 59
[0.033]
[0.016]
Phone -0.003
0.088
-0.01
[0.021]
Gender (Men=1)
[0.027]**
[0.011]
Electricity 0.019
0.014
-0.05
[0.088]
Speak indigenous language
[0.041]
[0.012]**
Assets
House were they live 0.004
Grades of schooling
.one
0.025
-0.011
[0.026]
[0.043]
[0.019]
Other houses -0.004
-0.018
0.014
[0.022]
Incomplete Elementary School
[0.034]
[0.017]
Electronic devices -0.074
0.039
0.019
[0.052]
Incomplete Junior High School
[0.031]
[0.013]
Bicycle -0.004
[0.019]
Size of locality
Between 15 and 100 thousand
-0.012
-0.007
Vehicle -0.002
habitants
[0.031]
[0.016]
[0.021]
-0.063
0.021
Between 2.5 and 15 thousand
Marital Status
habitants
[0.028]*
[0.016]
Domestic Partnership -0.041
[0.053]
-0.011
0.009
Less than 2.5 thousands
[0.024]
[0.012]
Divorce 0.013
habitants
-0.011
-0.017
[0.099]
Number of rooms for sleeping
[0.010]
[0.005]**
Separate -0.063
Number of children
-0.008
0.004
[0.059]
[0.006]
[0.003]
Married
-0.05
A member is a Oportunidades
-0.025
-0.004
[0.055]
beneficiary
[0.027]
[0.012]
Single -0.059
Medical insurance
[0.052]
IMSS
-0.011
-0.006
[0.019]
[0.009]
Seguro Popular
-0.002
0.041
[0.042]
[0.023]
Observations
2277
2
Start
being
obese
0.006
[0.014]
0.003
[0.012]
-0.004
[0.010]
0.032
[0.029]
-0.002
[0.012]
-0.003
[0.010]
-0.016
[0.020]
-0.021
[0.009]*
0.025
[0.010]*
0.023
[0.033]
-0.025
[0.042]
-0.004
[0.033]
0.004
[0.027]
-0.013
[0.028]
5340
Excludes pregnant women in 2002, 2005 or in between. Excluded categories: more than junior high, more than 100
thousand habitants, widow, and BMI lower than 25.
42
43
Table 14. Change in mental health indicators between obesity groups. MxFLS 2002 & 2005.
2002
Total
% Obese
% Obese in
in 2002 &
2002 &
not in
2005
2005
Change
% Not
obese in
2002 &
obese in
2005
% Not
obese in
2002 &
2005
Total
% Obese % Not obese
% Obese in
in 2002 & in 2002 &
2002 &
not in
obese in
2005
2005
2005
% Not
obese in
2002 &
2005
Percentage of the sample
100%
22%
7%
8%
63%
100%
22%
7%
8%
63%
Dying wishes
Feel fear
10.6%
30.8%
11.8%
33.5%
11.7%
31.3%
9.2%
27.8%
10.2%
30.2%
-0.004
-0.085
-0.005
-0.077
-0.008
-0.105
0.012
-0.056
-0.005
-0.090
Lack of confidence
Irritated
26.1%
35.8%
27.0%
38.1%
25.2%
34.1%
26.1%
35.3%
25.9%
35.3%
-0.050
-0.085
-0.045
-0.079
-0.043
-0.102
-0.046
-0.066
-0.053
-0.088
Useless
Loneliness
21.0%
28.1%
22.9%
30.2%
21.5%
25.3%
19.5%
24.4%
20.5%
28.2%
-0.027
-0.051
-0.023
-0.054
-0.012
-0.023
-0.024
-0.025
-0.031
-0.056
Nervous
Feel obsessive
42.2%
27.8%
44.4%
28.3%
44.1%
29.2%
41.8%
26.8%
41.3%
27.6%
-0.097
-0.064
-0.089
-0.054
-0.135
-0.103
-0.097
-0.043
-0.095
-0.065
Sadness
Tired
43.8%
44.7%
47.1%
49.2%
44.4%
47.0%
45.2%
46.0%
42.3%
42.7%
-0.036
-0.073
-0.047
-0.090
-0.028
-0.093
-0.071
-0.094
-0.029
-0.062
Lack of sexual interest
Pessimistic
24.1%
27.9%
28.1%
29.4%
29.5%
26.6%
22.1%
28.0%
22.3%
27.5%
-0.032
-0.055
-0.030
-0.044
-0.090
-0.050
-0.028
-0.068
-0.026
-0.057
Badly sleep
Spiritless
41.2%
37.8%
42.3%
39.5%
40.5%
39.4%
40.3%
37.6%
41.0%
37.1%
-0.030
-0.057
-0.010
-0.051
-0.037
-0.075
-0.021
-0.053
-0.038
-0.058
44
Table 15. Placebo regressions, impact of mental health indicators in change in weight.
Total
Sadness
Badly sleep
Spiritless
Feel obsessive
Lack of sexual interest
Nervous
Tired
Pessimistic
Irritated
Lack of confidence
Useless
Feel fear
Dying wishes
Loneliness
Observations
1
-0.175
[0.198]
0.128
[0.195]
-0.255
[0.200]
-0.148
[0.212]
-0.316
[0.246]
0.079
[0.196]
0.204
[0.196]
0.036
[0.214]
-0.162
[0.200]
0.135
[0.219]
-0.001
[0.237]
-0.077
[0.209]
-0.324
[0.311]
0.029
[0.220]
7865
Men
2
-0.068
[0.399]
-0.374
[0.371]
0.598
[0.410]
-0.535
[0.409]
0.109
[0.417]
-0.75
[0.411]
-0.19
[0.388]
0.24
[0.441]
0.678
[0.393]
-0.592
[0.474]
-0.156
[0.476]
0.318
[0.424]
0.277
[0.577]
-0.676
[0.441]
7108
1
0.067
[0.347]
0.423
[0.334]
-0.251
[0.359]
0.095
[0.376]
0.088
[0.485]
0.178
[0.343]
0.272
[0.339]
0.261
[0.396]
0.142
[0.355]
0.19
[0.415]
0.168
[0.452]
0.349
[0.383]
0.256
[0.711]
0.294
[0.453]
3287
2
0.426
[0.614]
-0.871
[0.561]
1.008
[0.641]
-1.128
[0.630]
-0.759
[0.734]
-0.433
[0.622]
-0.057
[0.581]
0.004
[0.691]
0.804
[0.603]
-0.413
[0.767]
-0.014
[0.779]
-1.338
[0.685]
-0.403
[1.156]
0.321
[0.771]
3071
Women
1
2
-0.327
-0.366
[0.236]
[0.526]
-0.085
-0.004
[0.236]
[0.495]
-0.244
0.397
[0.236]
[0.535]
-0.276
-0.295
[0.251]
[0.539]
-0.525
0.501
[0.276]
[0.513]
0.037
-0.843
[0.236]
[0.547]
0.171
-0.268
[0.237]
[0.521]
-0.054
0.394
[0.249]
[0.574]
-0.346
0.537
[0.238]
[0.518]
0.125
-0.625
[0.252]
[0.606]
-0.058
-0.409
[0.271]
[0.604]
-0.309
1.244
[0.243]
[0.543]*
-0.476
0.347
[0.333]
[0.675]
-0.052
-1.234
[0.245]
[0.546]*
4578
4037
Note: In the first column each mental health indicator was related to weight in a separate regression. The
second column it is just one regression. The controls used were: dummies by age group, gender,
indigenous status, dummies by schooling, size of locality, number of rooms, families with
Oportunidades, number of children, phone, television, other possessions, marital status and the intercept.
Standard errors in brackets. * significant at 5%; ** significant at 1%
45
Table 16. Impact of Obesity and Weight on Mental Health indicators. Fixed Effects. By size of community and gender
Women
Men
Total
Rural
Urban
Rural
Urban
Obesity
Weight
Obesity
Weight
Obesity
Weight
Obesity
Weight
Obesity
Weight
Sadness
-0.018
-0.001
0.008
0
-0.041
0
-0.017
-0.001
-0.015
0
[1.04]
[0.66]
[0.22]
[0.14]
[1.38]
[0.13]
[0.42]
[0.80]
[0.41]
[0.31]
Badly sleep
0.007
-0.001
0.032
0
0.015
0
-0.035
-0.003
0.022
-0.001
[0.37]
[1.09]
[0.94]
[0.11]
[0.47]
[0.18]
[0.88]
[1.98]*
[0.54]
[0.31]
Spiritless
0.012
-0.001
-0.001
-0.001
0.013
-0.001
-0.038
-0.003
0.086
0.001
[0.68]
[1.42]
[0.04]
[0.45]
[0.44]
[0.49]
[0.97]
[2.11]*
[2.40]*
[0.80]
Feel obsessive
0.027
-0.001
0.024
0
0.03
0
0.039
-0.002
0.031
0
[1.70]
[0.95]
[0.75]
[0.24]
[1.08]
[0.04]
[1.06]
[1.36]
[0.90]
[0.10]
Lack of sexual interest
0.03
0
0.019
0.001
0.065
0.001
0.017
0
0.011
-0.001
[1.71]
[0.35]
[0.52]
[0.84]
[2.01]*
[0.89]
[0.45]
[0.09]
[0.35]
[0.63]
Nervous
0.016
-0.002
0.007
-0.002
0.01
-0.002
0.036
-0.003
0.03
-0.002
[0.94]
[3.06]**
[0.21]
[1.29]
[0.34]
[1.56]
[0.90]
[1.83]
[0.80]
[1.18]
Tired
0
-0.003
0.02
-0.001
0.004
-0.005
-0.018
-0.002
-0.013
-0.003
[0.00]
[3.88]**
[0.59]
[0.63]
[0.12]
[3.25]**
[0.42]
[1.19]
[0.34]
[1.78]
Pessimistic
-0.012
-0.002
-0.073
-0.003
0.036
-0.001
-0.046
-0.002
0.009
-0.001
[0.78]
[2.55]*
[2.26]*
[1.77]
[1.30]
[0.98]
[1.30]
[1.66]
[0.28]
[0.87]
Irritated
0.018
0
-0.006
-0.001
0.05
0.002
-0.005
0
0.009
-0.002
[1.04]
[0.40]
[0.18]
[0.78]
[1.71]
[1.11]
[0.12]
[0.17]
[0.26]
[1.17]
Lack of confidence
-0.002
-0.002
-0.039
-0.002
0.014
-0.002
-0.01
-0.002
0.016
-0.001
[0.14]
[2.77]**
[1.26]
[1.27]
[0.50]
[1.93]
[0.30]
[1.49]
[0.52]
[0.59]
Useless
-0.005
-0.001
-0.046
-0.001
0.029
0.001
-0.008
-0.002
-0.007
-0.002
[0.32]
[1.12]
[1.54]
[0.75]
[1.12]
[1.02]
[0.24]
[1.56]
[0.26]
[1.42]
Feel fear
0.022
-0.002
-0.022
-0.001
0.057
-0.001
-0.012
-0.003
0.05
-0.003
[1.40]
[2.45]*
[0.67]
[0.76]
[2.02]*
[0.46]
[0.33]
[2.00]*
[1.50]
[2.03]*
Dying wishes
0.011
0
0.003
0
0.013
0
0.028
0
0.011
-0.001
[1.02]
[0.41]
[0.14]
[0.39]
[0.63]
[0.12]
[1.22]
[0.05]
[0.57]
[0.90]
Loneliness
0.003
-0.001
-0.007
-0.002
-0.001
0
-0.013
-0.001
0.033
-0.001
[0.22]
[1.50]
[0.22]
[1.11]
[0.04]
[0.23]
[0.39]
[0.70]
[1.22]
[0.68]
Observations
21577
21725
5453
5494
7265
7307
3849
3881
4945
4977
Number of groups
13202
13232
3127
3137
4355
4361
2451
2457
3289
3297
Each impact is estimated in a separate regression. The controls used were: dummies by age group, gender, indigenous status, dummies by schooling,
size of locality, number of rooms, families with Oportunidades, number of children, phone, television, other possessions, marital status, and
intercept. t-statistics in bracket. * significant at 5%; ** significant at 1%
46
Table 17. Impact of Obesity and Weight on Mental Health indicators only Urban Women. Fixed Effects. By years
of schooling.
Incomplete
More
None
Incomplete Primary
Secondary
Obesity
Weight
Obesity
Weight
Obesity
Weight
Obesity
Weight
Sadness
-0.146
-0.001
-0.074
0.007
-0.045
-0.002
-0.08
0
Badly sleep
[1.11]
-0.229
[0.23]
-0.007
[0.86]
0.005
[1.66]
0.009
[1.10]
0.046
[0.99]
0
[0.83]
0.031
[0.07]
-0.003
Spiritless
[1.82]
-0.127
[1.18]
-0.001
[0.06]
-0.059
[2.05]*
0
[1.08]
0.027
[0.19]
-0.001
[0.31]
0.007
[0.71]
0.001
Feel obsessive
[0.96]
-0.083
[0.10]
0
[0.71]
0.037
[0.07]
0.006
[0.61]
0.058
[0.52]
-0.001
[0.07]
-0.037
[0.29]
-0.001
Lack of sexual interest
[0.65]
-0.088
[0.02]
0.011
[0.46]
-0.011
[1.52]
0.005
[1.50]
0.106
[0.65]
0
[0.46]
0.041
[0.29]
0.004
Nervous
[0.42]
-0.247
[1.24]
-0.006
[0.10]
-0.024
[0.80]
0.006
[2.41]*
0.061
[0.15]
-0.001
[0.50]
0.031
[1.00]
-0.004
Tired
[1.82]
-0.127
[0.96]
-0.011
[0.27]
0.037
[1.34]
0.001
[1.47]
-0.007
[0.54]
-0.005
[0.33]
0.015
[0.95]
0.003
Pessimistic
[0.91]
-0.132
[1.72]
0
[0.43]
0.061
[0.29]
0.006
[0.17]
0.019
[2.55]*
-0.002
[0.15]
0.055
[0.77]
-0.004
Irritated
[1.09]
0.115
[0.06]
0.003
[0.76]
-0.035
[1.46]
0.003
[0.49]
0.042
[1.30]
0.003
[0.68]
0.056
[1.05]
-0.001
Lack of confidence
[0.87]
-0.125
[0.53]
-0.006
[0.40]
-0.011
[0.72]
0.001
[1.03]
0.015
[1.38]
-0.002
[0.64]
0
[0.35]
-0.002
Useless
[1.06]
0.154
[1.04]
0.003
[0.14]
0
[0.27]
0.008
[0.40]
0
[1.09]
0
[0.01]
0.065
[0.69]
0
Feel fear
[1.34]
-0.104
[0.53]
-0.002
[0.00]
0.067
[1.90]
0.002
[0.00]
0.05
[0.03]
-0.002
[0.93]
0.052
[0.02]
-0.002
Dying wishes
[0.86]
0.154
[0.32]
-0.002
[0.83]
-0.008
[0.56]
-0.001
[1.29]
0.002
[1.12]
0
[0.62]
0.086
[0.44]
0.003
Loneliness
[1.41]
-0.191
[0.33]
-0.001
[0.13]
0.043
[0.16]
0.007
[0.06]
-0.002
[0.33]
-0.003
[1.57]
0.029
[1.06]
0
Observations
[1.39]
570
[0.19]
571
[0.53]
1164
[1.62]
1171
[0.04]
3974
[1.39]
3997
[0.34]
1501
[0.01]
1512
Number of groups
398
399
795
797
2571
2575
1070
1074
Each impact is estimated in a separate regression. The controls used were: dummies by age group, gender,
indigenous status, dummies by schooling, size of locality, number of rooms, families with Oportunidades, number
of children, phone, television, other possessions, marital status, and intercept. t-statistics in bracket. * significant at
5%; ** significant at 1%
47
Table 18. Impact of Obesity and Weight on Mental Health indicators only Urban Women. Fixed Effects. By age.
20 to 29
30 to 39
40 to 49
50 to 60
Obesity
Weight
Obesity
Weight
Obesity
Weight
Obesity
Weight
Sadness
-0.07
[0.84]
-0.005
[1.43]
-0.13
[1.87]
0.003
[0.75]
0.014
[0.19]
0.002
[0.50]
0.008
[0.09]
-0.002
[0.39]
Badly sleep
-0.008
[0.09]
-0.003
[0.90]
0.085
[1.14]
0.005
[1.39]
0.141
[1.82]
0.006
[1.46]
0.026
[0.29]
0.002
[0.34]
Spiritless
0.061
[0.75]
-0.003
[0.86]
-0.008
[0.11]
0.006
[1.81]
0.079
[1.03]
-0.001
[0.13]
0.068
[0.81]
-0.001
[0.29]
Feel obsessive
0.025
[0.33]
-0.005
[1.69]
0.011
[0.16]
-0.002
[0.55]
0.13
[1.84]
0.004
[1.04]
0.073
[0.98]
0.006
[1.39]
Lack of sexual interest
0.159
[2.21]*
0.002
[0.67]
0.048
[0.67]
0
[0.09]
-0.033
[0.36]
0.006
[1.23]
0.103
[0.81]
-0.01
[1.15]
Nervous
0.03
[0.35]
0
[0.04]
-0.028
[0.40]
-0.005
[1.54]
0.023
[0.30]
-0.001
[0.27]
-0.092
[1.01]
-0.004
[0.68]
Tired
0.1
[1.15]
-0.003
[0.98]
-0.064
[0.86]
-0.005
[1.36]
0.111
[1.48]
-0.001
[0.17]
-0.043
[0.49]
-0.006
[1.18]
Pessimistic
0.123
[1.55]
-0.001
[0.41]
-0.019
[0.29]
-0.001
[0.44]
0.087
[1.31]
0.002
[0.70]
0.134
[1.61]
0.001
[0.19]
Irritated
0.019
[0.24]
-0.002
[0.47]
0.052
[0.75]
-0.001
[0.22]
0.091
[1.24]
0.009
[2.44]*
-0.137
[1.63]
-0.009
[1.88]
Lack of confidence
0.004
[0.05]
-0.009
[2.91]**
-0.06
[0.93]
-0.004
[1.13]
0.107
[1.55]
0.006
[1.68]
-0.069
[0.92]
-0.004
[0.81]
Useless
0.008
[0.12]
-0.005
[1.72]
0.04
[0.67]
-0.001
[0.46]
0.013
[0.20]
0.005
[1.39]
-0.001
[0.01]
-0.001
[0.23]
Feel fear
0.014
[0.18]
-0.004
[1.15]
-0.001
[0.01]
0.001
[0.38]
0.106
[1.51]
0.008
[2.14]*
0.061
[0.70]
-0.003
[0.66]
0.144
[2.73]**
0.002
[0.78]
0.02
[0.43]
-0.001
[0.53]
0.089
[1.70]
0.005
[1.85]
-0.055
[0.87]
-0.001
[0.41]
Loneliness
0.016
[0.20]
-0.003
[0.99]
-0.043
[0.65]
-0.002
[0.55]
-0.033
[0.46]
0.006
[1.70]
-0.077
[0.90]
-0.001
[0.16]
Observations
Number of groups
1815
1296
1828
1301
2235
1624
2250
1630
1806
1312
1813
1315
1176
857
1181
859
Dying wishes
Each impact is estimated in a separate regression. The controls used were: dummies by age group, gender,
indigenous status, dummies by schooling, size of locality, number of rooms, families with Oportunidades,
number of children, phone, television, other possessions, marital status, and intercept. t-statistics in bracket. *
significant at 5%; ** significant at 1%
48
49