Food consumption, food expenditure, anthropometric status and nutrition-related diseases in Mexico

S. Barquera, C. Hotz, J. Rivera, L. Tolentino, J. Espinoza, I.Campos and T. Shamah, National Institute of Public Health, Cuernavaca, Mexico

Introduction

Mexico, as are other Latin American countries, is experiencing an epidemiological and nutrition transition characterized by a rapid rise in the prevalence of obesity and chronic diseases such as diabetes mellitus, high blood pressure and cardiovascular diseases (CVDs) (Flores et al., 1998; Rivera-Dommarco et al., 2001). Although the morbidity and mortality related to acute communicable diseases, such as diarrhoeas and respiratory infections, undernutrition and some micronutrient deficiencies, have shown important reductions during the last three decades, different forms of undernutrition, such as child stunting and anaemia, remain relevant public health problems.

Mexico’s gross national product (GNP) per capita increased by 62 percent between 1970 and 2003. This change is greater than that observed for the Latin America and the Caribbean (LAC) region as a whole (43 percent) (Figure 1a). Currently, 75 percent of the population live in urban locations. The percentage of population living in rural locations decreased by 39 percent in the 1970 to 2003 period (Figure 1b). During recent decades, health status has been affected by environmental and economic conditions. Child mortality has decreased by 74.5 percent in 33 years, which is similar to the LAC regional average (Figure 1c) but three times greater than the rate observed for other Organisation for Economic Co-operation and Development (OECD) countries during the same period. In Mexico, health expenditure as a percentage of GNP was lower than the average for Latin American countries during the 1998 to 2002 period.

Currently, Mexico has more than 100 million inhabitants. As one of the world’s mega-countries, implementation of social policies is complex, and the expected impact difficult to measure (Barquera, Rivera-Dommarco and Gasca-Garcia, 2001). Although Mexico has been developing food and nutrition programmes since the early twentieth century and has devoted more resources to these programmes than any other Latin American country (Rivera-Dommarco et al., 2001; Rivera et al., 2004b), until recently social policies were not evaluated (Barquera, Rivera-Dommarco and Gasca-Garcia, 2001; Ministry of Social Development, 2000). Food and nutrition programmes in the last eight years have improved their targeting and have incorporated objectives related to improving nutritional status and education. It is expected that these improved programmes will have positive impacts on nutrition and health. Conversely, consumer access to industrialized foods has increased in recent years owing to technological developments and economic growth.

FIGURE 1
Development indicators in Mexico, Latin America and the Caribbean

Mexico is also characterized by major epidemiological differences across country regions, urban/rural residence and socio-economic status. Imbedded in these differences is the polarization of the transition, which means that different subpopulations within the country are undergoing different stages of transition. In Mexico, polarization has been described across four regions (Bobadilla et al., 1993; Hernandez-Diaz et al., 1999; Rivera-Dommarco et al., 2001). The North region is the most industrialized. It has a higher per capita income and infrastructure level than the rest of the country, close cultural and economic relations with the border states of the United States and adequate access to basic and health services. The Central region is less developed than the North, but still contains large developed cities such as Guadalajara and many rural towns that live from agriculture. Mexico City is the third most populated city in the world. It has a very developed economy and access to food and basic services, combined with high immigration from the south and poverty pockets. It also receives several subsidies and social programmes targeting the poor. The South region is considered the least developed. It has the largest rural and indigenous population in the country, and access to certain basic services and subsidies is limited; in this region health problems such as infectious diseases and undernutrition still represent a relevant public health concern. While the health sector in all of these regions is still facing the challenge of preventing and treating acute diseases (Frenk et al., 1991), non-communicable chronic diseases such as diabetes and hypertension are rapidly increasing.

In order to address this complex situation, it will be important to understand how nutrition and related health conditions have evolved together over the last decades in the context of social, economic and market changes. The following subsections document the main characteristics of the epidemiologic and nutrition transition in Mexico, including nutrient intake, trends in food expenditure, nutritional status, prevalence of obesity and chronic diseases and mortality trends, with emphasis on polarization among the different regions of the country, between urban and rural populations and among socio-economic groups. The information is used to discuss the double burden of disease and the role of current and future national programmes in addressing emerging health challenges.

Methods

The analysis in this case study is based on diverse nationwide databases generated from surveys collected mostly by the National Institute of Public Health and from other cross-sectional surveys and registries collected by diverse governmental agencies (Table 1). Variables and measurements were stratified by region, urban/rural residence and socio-economic status, when possible. The case study team also analysed trends over time for variables such as nutrient intake in women, household food expenditure, prevalence of chronic diseases and mortality.

TABLE 1
Socio-demographic and health surveys used in this report

Country regions

The country was divided into four regions with common geographic and socio-economic characteristics: 1) North region, which comprises Baja California, Southern Baja California, Coahuila, Durango, Nuevo Leon, Sonora, Sinaloa, Tamaulipas and Zacatecas; 2) Central region, which comprises Aguascalientes, Colima, Guanajuato, Hidalgo, Jalisco, Mexico, Michoacan, Nayarit, Querétaro, San Luis Potosí and Tlaxcala; 3) Mexico City; and 4) South region, which comprises Campeche, Chiapas, Guerrero, Morelos, Oaxaca, Puebla, Quintana Roo, Tabasco, Veracruz and Yucatan. This regionalization scheme was used in epidemiologic transition analysis for intra-country comparisons (Hernandez-Diaz et al., 1999; Rivera et al., 2002; Sepulveda-Amor et al., 1990).

Place of residence

In addition, subjects were classified as urban or rural according to the population size of their place of residence. For variables obtained from the Mexican Nutrition Survey (MNS), these classifications were ³ 2 500 inhabitants for urban areas and < 2 500 inhabitants for rural ones, while for variables obtained from the Mexican Health Survey (MHS) and the Mexican Household Income and Expenditure Survey (MHIES) the corresponding figures were ³ 15 000 inhabitants and < 15 000 inhabitants.

Socio-economic status index

MNS-1 (1988) and MNS-2 (1999) obtained socio-economic information such as household conditions, basic services infrastructure (i.e., water source and disposal) and possession of domestic appliances (i.e., radio, television and refrigerator). A principal components factor analysis was carried out using this information to extract a main factor that explained more than 50 percent of the socio-economic information variability (Flores et al., 1998; Long-Dunlap et al. 1995). This factor was divided into quintiles and used as a relative measure of socio-economic status.

Nutrient and dietary information

Nutrient intake information for Mexico was obtained from MNS-1 and MNS-2, which are two nationwide representative surveys with sufficient sampling power to allow data to be disaggregated by region and urban/rural location. Detailed descriptions of the survey sampling procedures and methods have been published elsewhere (Resano-Perez et al., 2003; Rivera-Dommarco et al., 2001; Sepulveda-Amor et al., 1990). Food and nutrient intake information was available for females aged 12 to 49 years and for children aged one to 11 years in 1999. These surveys also had anthropometric information for women aged 12 to 49 years and for children under five years of age.

A 24-hour dietary recall (24HDR) was administered to obtain dietary information for randomly selected sub-samples of 9 449 (MNS-2) and 2 596 (MNS-1) women. Trained and standardized personnel applied the questionnaires and converted consumption into grams or millilitres of food items. For MNS-2, aberrant food consumption was reviewed by hand and updated when a clear mistake was detected, or eliminated if the value was not biologically plausible. Complete food intake data were not available from MNS-1, but a database with estimated nutrient intake generated from 24HDR was obtained and used for the analysis. Nutrient intakes were then estimated for foods, using a comprehensive nutrient composition database compiled from several sources (Muñoz et al., 1996; Souci, Fachmann and Kraut, 2000; USDA/ARS, 1999; Wuleung and Flores, 1962). Nutrient adequacies were calculated relative to the dietary reference intakes (Institute of Medicine, 2000a; 2000b; 2001; 2002; National Research Council, 1989); the estimated average requirement was used for protein, vitamins A and B12, folic acid, iron and zinc; the adequate intake was used for calcium and vitamin D; and the estimated energy requirement was used for energy. A cut-off of 30 percent of total energy intake was considered fat adequacy. Requirements were adjusted for pregnant and lactating women. To estimate intake by food groups, grams consumed were aggregated into selected groups and reported as mean consumption, stratified by region, location and socio-economic index tertile.

Estimation of household adult equivalent

To estimate household food expenditure using MHIES, the individual adult equivalent (AE) was obtained by dividing the recommended dietary allowance (RDA) for energy of each household member - according to age and sex - by the energy RDA for an average adult. The sum of all of the individual AEs within a household was then computed to obtain the number of household AEs. Family members not currently living in the house were excluded from this estimation, but their income contribution was included in the total household income.

Quantity of food purchased per AE

Household food quantities purchased per AE were used as a proxy for food intake. Data from seven MHIES conducted between 1989 and 2002 were used. These surveys were collected by the National Institute of Geography, Informatics and Statistics. They contain nationally representative information on approximately 15 000 households and their members. Information was available for 1989, 1992, 1994, 1996, 1998, 2000 and 2002. A questionnaire based on a seven-day diary record of house purchases and containing about 215 foods was included in each MHIES, with slight variations in some years. A database with the quantities of each food purchased per household was developed. All foods were converted into grams, and food group quantities were estimated for cereals, meats, eggs, milk and dairy products, legumes, fats and oils, vegetables, fruits, and sugars. In addition, soda, alcohol and tobacco were converted into millilitres and grams and included in the final tables. From this database, the following indicators were obtained: 1) percentage of families purchasing each food group during the seven-day survey period; 2) percentage of total food expenditure allocated to each food group; 3) mean g/day per AE in families reporting expenditure on that item; and 4) per capita mean g/day per AE. The data were stratified by region, urban/rural residence and income quintile (only the extreme income quintiles are presented in this case study).

Percentage of food expenditure outside the home

In addition to food expenditure at home, MHIES also reported expenditure on food outside the home. It was therefore possible to calculate food expenditure outside the home as a percentage of total food expenditure per AE. This expenditure was not disaggregated according to food item.

Breastfeeding characteristics in Mexico

The information on breastfeeding practices in Mexico in this case study is based on a previous analysis of MNS-2 (González-Cossío et al., 2003).

Children’s nutritional status

Nutritional status derived from anthropometric indicators was available for children under five years of age from MNS-1 and MNS-2. In addition, MNS-2 collected information for children of five to 12 years of age. Using weight, height (or length) and age, the case study team calculated height-for-age, weight-for-height and weight-for-age Z-scores relative to the National Center for Health Statistics/World Health Organization (NCHS/WHO) reference population (HAZ, WAZ and WHZ, respectively) (WHO, 1983; 1995). Underweight was defined as WAZ < -2, stunting as HAZ < -2, wasting as WHZ < -2 and overweight as WAZ > +2. For school-age children, MNS-2 used two different criteria of classification for overweight (Cole et al., 2000; Must, Dallal and Dietz, 1999).

Adult nutritional status and chronic disease prevalence

Adult nutritional status and prevalence rates of chronic diseases were obtained from two representative national surveys: the Mexican Chronic Diseases Survey (MCDS, 1994) and MHS (2000) (Olaiz, Rojas and Barquera, 2003). MCDS was implemented in four urban areas of Mexico and is representative of the country’s urban locations. MHS was implemented in urban and rural locations with sample power to be representative of each state. In this survey, the rural population was defined as those living in locations with < 15 000 inhabitants. Prevalence rates of low weight (body mass index [BMI] < 18.5 kg/m²), normal weight (BMI 18.5 to 24.9 kg/m²), overweight (BMI > 25 kg/m²) and obesity (BMI > 30 kg/m²) were calculated. Diabetes mellitus was defined as a fasting glucose concentration of ³ 126 mg/dl, or a non-fasting concentration of > 200 mg/dl, and/or a previous medical diagnosis of diabetes (Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, 2005; WHO, 1994). Hypertension was defined as systolic pressure ³ 140 mm/Hg and/or diastolic pressure ³ 90 mm/Hg for people under 60 years of age. For those aged 60 years and over, hypertension was defined as systolic pressure ³ 160 mm/Hg and/or diastolic pressure ³ 90 mm/Hg. Hypercholesterolaemia was defined as fasting blood cholesterol concentrations of ³ 200 mg/dl, and hypertriglyceridaemia as fasting triglyceride concentrations of ³ 150 mg/dl (Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults, 2001).

Trends in communicable and non-communicable disease mortality

The mortality indicators for the years 1950 to 2000 were obtained from estimations made by the Ministry of Health for the national health programme (Ministry of Health, 2001). Mortality rates for intestinal infectious diseases, acute respiratory diseases and undernutrition in children aged one to nine years, and ischaemic heart disease (IHD) and diabetes mellitus in adults aged 20 to 74 years were obtained from the National Institute of Informatics, Statistics and Geography (INEGI) using the International Classification of Diseases version 9 (ICD-9) for the years 1979 to 1995 (WHO/PAHO, 1978) and version 10 (ICD-10) for 2000 and 2002 (WHO, 1992).

TABLE 2
Women's energy and nutrient intakes and percentage adequacies by region and location, 1988 and 1999

^{a, b, c, d} Different superindices represent statistically significant differences among regions. º Statistically different from rural locations.
Sources: MNS-1, 1988; MNS-2, 1999.

TABLE 3
Women’s energy and nutrient intakes and percentage adequacy, by socio-economic status and education, 1988 and 1999

Sample sizes: children < 11 years of age, 1 249; females 12 to 49 years of age - 1988, 9 449; 1999, 2 596.
^a,b,c Different superindices represent statistically significant differences among socio-economic quintiles.
* Statistically different from middle and higher education.
Sources: MNS-1; MNS-2.