Design and sample

Ensanut-MC 2016 and Ensanut 2012 are probabilistic population-based surveys with a multistage, stratified sampling design, representative at the national and regional levels and for rural and urban areas. Detailed sampling procedures for both surveys are described elsewhere.^12,13 Briefly, Ensanut-MC 2016 was conducted between May and September of 2016 and collected information from 9 479 households with a household response rate of 77.9%.¹² Ensanut 2012 was conducted between October 2011 and May 2012, and collected information from 50 528 households, with a household response rate of 87%.¹³ Detailed dietary information was obtained for random subsamples in both surveys using 24-hour recalls (n=4 341 for Ensanut-MC 2016 and n=10 886 for Ensanut 2012). For both surveys, a second 24-hour recall on a non-consecutive day was collected for a subsample (»7% for Ensanut-MC 2016 and »9%). For the present analysis, we only used the first 24-hour recall and excluded children younger than 5y (n=3 190), pregnant and lactating women (n=245), and individuals with implausible intake (n=152). Plausible intake was defined as energy intakes between ±3 standard deviations from the ratio of energy intake and energy requirement, estimated using the Institute of Medicine equations for body weight maintenance. A thoroughly detailed description of the methods used to identify implausible intake is described elsewhere.^14,15 Thus, the study sample included school-aged children (5-11y), adolescent (12-19y) and adult (≥20y) men, and non-pregnant, non-lactating adolescent and adult women with complete socioeconomic information (n=3 661 for Ensanut-MC 2016 and n=7 979 for Ensanut 2012).

Dietary assessment

The 24-hour recall was collected through an interview to ≥15-year old individuals by trained personnel between Monday and Sunday, using an automated 5-step multiple-pass method.¹⁶ Participants were asked to report all foods and beverages consumed the previous day. The person responsible for food preparation in the household provided information of food intake of children and adolescents <15y old. Children or adolescents complemented the interview by reporting food eaten away from home. Intake could be reported as individual foods or beverages (e.g. apple or soda) or mixed dishes (e.g. sandwich or smoothie). Mixed dishes could be then disaggregated to its ingredients if the participant knew the amounts used in its preparation. Otherwise, the mixed dish was recorded as a standard preparation. Tortilla and other typical foods of specific regions were weighted to capture the variability between different regions of the country. The energy content was estimated using the 2016 food composition table compiled by the INSP.^*

Food group classification

Foods reported as disaggregated and standard recipes were analyzed at their ingredient level, while beverages were kept as a single unit. For example, a sandwich was disaggregated to bread, cheese, ham, avocado, tomato, etc., whereas beverages prepared with different ingredients (e.g. coffee with sugar) were maintained as a single unit. Milk and dairy beverages were the exception and were treated as disaggregated foods in order to avoid underestimating milk intake.

Foods and beverages were then classified into 24 specific food groups, which in turn were aggregated into eight major food groups as follows: 1) whole grains, refined grains and tubers, grouped as ‘cereals and tubers’; 2) legumes, seeds and nuts, grouped as ‘legumes and nuts’; 3) low-fat dairy, high-fat dairy, and dairy with added sugar, grouped as ‘milk and dairy’; 4) eggs, red meat, processed meat, poultry, and fish and seafood, grouped as ‘meat and animal products’, 5) fruits, vegetables, and 100% fruit or vegetable juices, grouped as ‘fruits and vegetables’, 6) fats and oils, 7) manufactured sugar-sweetened beverages (SSBs), homemade SSBs and alcoholic beverages, grouped as ‘SSBs’, and 8) grains with added sugar, sweets and sweeteners, desserts, and snacks, grouped as ‘products high in saturated fat and/or added sugar (HSFAS)’. The energy contribution from the abovementioned food groups accounted for 99.0% of the total energy intake. Foods and beverages not included were condiments, baby foods, beverages used for cooking (beer or juices), and non-dairy non-sweetened beverages.

Whole grains were defined as grains with ≥10g of total fiber per 100g of carbohydrates, and refined grains were classified otherwise.¹⁷ Grain products with ≥13% of total energy from saturated fat or added sugar, regardless of their fiber content, were classified into the HSFAS products group. Low-fat dairy included milk and dairy with <3.5g of fat per 100 kcal, while high-fat dairy included milk and dairy with ≥3.5g of fat per 100 kcal. Dairy with ≥10g of sugar per 100 kcal was classified as dairy with added sugar, regardless of the fat content. The SSBs group included non-dairy beverages with added sugar and all alcoholic beverages. Within the SSBs group, manufactured SSBs included soft drinks, fruit beverages, flavored water, iced teas, soy drinks, and sports and energy drinks; and homemade SSBs included aguas frescas (traditional Mexican beverages usually prepared with water, fruit, and sugar), atole (a corn meal beverage) prepared with water and sugar, and homemade coffee or tea with sugar. The HSFAS products group included processed non-basic foods with 13% or more of total energy from saturated fat and/or added sugars. The HSFAS products group included salty snacks, grain-based desserts, ready-to-eat cereals, and other processed products with >13% of total energy from saturated fat and/or added sugars. The 13% cutoff point for saturated fats and/or added sugar was based on the International Choices Program.¹⁸

Given that food groups 1 through 6 contribute to the intake of essential nutrients, they were considered basic foods; whereas groups 7 and 8 (SSBs and HSFAS products), which provide a high amount of energy with low nutrient density, were considered discretionary foods.

Food groups recommendations

In order to document the adequacy of the population’s mean intake, we used the same recommendations as the previously mentioned analysis from 2012 based on the MDPAG.⁹ The MDPAG present thirteen specific and detailed dietary plans for different population groups that range from 1 000 kcal/d to 2 600 kcal/d, based on age and activity level. Each dietary plan suggests a set amount of daily servings for eight food groups (vegetables, fruits, cereals, legumes, meat and animal products, milk and dairy, sugars and fats). In order to compare intake against the dietary plans, we estimated the percentage of energy contribution (EC) from each food group for dietary plans included in the range between 1 300 and 2 600 kcal/day (nine dietary plans), corresponding to plans for school-aged children, adolescents and adults. Then, we estimated the average EC for each food group of the abovementioned plans. For the present analysis, the EC from vegetables and fruits was added. The MDPAG recommend decreasing intake of SSBs, pastries, cookies and desserts, and foods high in sugar, fat and sodium; and recommend a maximum 6-9% EC from added sugars. Nevertheless, they do not provide a maximum recommended EC for discretionary foods beyond added sugars. Thus, to be able to allocate products like salty snacks and cookies into a specific food group, this analysis combined non-essential foods and beverages into a “discretionary foods” category with a maximum EC of 13%, of which up to 3% can be supplied by SSBs and up to 10% by HSFAS products. The 3% cut-off point for SSBs was selected to be in line with the American Heart Association recommendation.¹⁹ This 13% maximum is not to be interpreted as a dietary guideline, but can be rather interpreted as a short-term dietary goal suggested by the authors of this paper, and could be further reduced in a longer term.

The recommended EC of food groups used in this analysis were as follows: 33% from cereals, 12% from legumes, 10% from milk and dairy, 10% from meat and animal products, 12% from fruit and vegetables, 10% from fats and oils, and no more than 3% from SSBs and 10% from HSFAS products.

Sociodemographic characteristics

The place of residence was categorized into rural and urban areas with locations of ≤2 500 inhabitants classified as rural, and otherwise, as urban. The geographical region was classified as North, Central, or South. The socioeconomic status (SES) index was assessed using principal components analysis based on household characteristics and assets; households were then categorized into tertiles. An individual was classified as having indigenous ethnicity if he/she lived in a household where the head of household or his/her spouse spoke an indigenous language. Education level (only for adults ≥20y) was categorized into four groups: lower than elementary school, finished elementary school, finished middle school, finished high school or higher (including normal and technical high school).

Statistical analysis

All analyses were conducted in Stata version 14.2^‡ and were weighted to be nationally representative and to account for the complex survey design. Means and 95% confidence intervals for the percent EC from the eight major food groups and 24 specific food groups were estimated for the overall population, whereas the EC from the eight major food groups and a key selection of specific food groups (whole grains, refined grains, red meat, processed meat, fruits, vegetables, manufactured SSBs and homemade SSBs) was stratified by population subgroups. Differences in EC among subgroups were compared using bivariate linear regression models. No further adjustments were made for multiple comparison, but p-values and confidence intervals are presented in all tables. In order to compare between the percent ECs of 2012 and 2016, we estimated multivariate linear regression models using pooled data from both surveys; estimating one model per food group and adjusted by age group, sex, residence area, geographical region, tertiles of socioeconomic level, indigenous ethnicity, and weekend days (Friday through Sunday). Given that education level was only relevant for adults, we favored including all age groups and excluded the education level as a covariate. Predicted means were estimated using Stata’s -margins- command.

Informed consent was obtained from individuals aged ≥18y, and from the father, mother, or guardian of participants aged <18y. Informed assent was obtained for children and adolescents (5-17y). Both surveys were approved by the Research, Biosafety, and Ethics Committees of the INSP in Cuernavaca, Mexico.

Major food groups in 2016

Overall, cereals had the highest percent of EC of the total daily energy intake (30.1%), followed by meat and animal products (15.9%), HSFAS products (15.5%), and SSBs (10.5%) (table II). Fats and oils had an EC of 8.7%, while milk and dairy contributed 8% to the total daily energy intake. Fruits and vegetables, and legumes had the lowest contribution to the total daily energy intake, with 6.4 and 3.8%, respectively. HSFAS products and SSBs, considered discretionary foods, contributed with 26% to the total daily energy intake. Legumes had the lowest intake relative to the recommendations, with 8.2 percentage points (pp) of EC below the percentage suggested by the MDPAG; followed by fruits and vegetables (5.6 pp lower than recommended). The highest intake relative to recommendations was for SSBs, being 7.5 pp above the maximum percent suggested. Intake of meat and animal products, and HSFAS products also diverged significantly from the recommendations (5.9 and 5.5 pp above, respectively). In other words, SSBs, meat and animal products and HSFAS were 250, 59 and 55% above the recommended intake, respectively, whereas legumes, and fruits and vegetables were 68 and 47% below the recommended intake.

There were important differences in EC among population subgroups. Within age groups, the average EC of the diet of adolescents and adults was lower in milk and dairy, and higher in SSBs, compared to school-aged children. Also, the EC percentages of cereals, legumes, and meat and animal products were higher in adults compared to school-aged children, whereas the EC of HSFAS products was higher in school-aged children compared to adults. Discretionary foods contributed 27.7% to total daily energy intake in school-aged children; 28.4%, in adolescents, and 25%, in adults; in other words, EC was the 14.7, 15.4 and 12 pp above the recommended intake, respectively.

Between men and women, the only observed difference was for SSBs, with a higher EC for men vs. women. The diets of individuals living in rural areas, in the Central and South regions of the country, of low-SES, of indigenous ethnicity, and who did not complete elementary school had higher EC from cereals and legumes, and lower EC for milk and dairy, meat and animal products, and HSFAS products relative to their respective counterparts (individuals living in urban areas, in the North region, of medium- and high-SES, of non-indigenous ethnicity, and who had completed elementary, middle, or high school). Exceptions being legumes, milk and diary, and HSFAS products that did not differ between regions, and HSFAS products that did not differ by indigenous ethnicity. Moreover, the diets of individuals of medium- and high-SES had a higher EC from fruits and vegetables, relative to individuals from low-SES. The percentage of EC from SSBs was higher in the North region compared to the Central and South regions; and lower among individuals of indigenous ethnicity compared to their counterparts.

Specific food groups in 2016

Among the cereals, whole grains had the highest EC, with 21.5% of total daily energy intake. In the meat and animal products group, red meat had the highest EC (6.4%), and fish and seafood, the lowest (0.7%). Within the SSBs group, manufactured SSBs contributed 6.0% to the total daily energy intake, followed by homemade SSBs, with 3.9%. Pastries and ready-to-eat cereals with added sugar had the highest EC within the HSFAS products group, with 9.4% of the total daily energy intake (table III).

Differences in EC between population subgroups were also observed for the specific food groups selected. Among age groups, the diets of adolescents and adults had a higher EC from whole grains and manufactured SSBs, and a lower EC from refined grains, compared to school-aged children (table IV). Individuals living in rural areas, in the Central and South regions of the country, from low-SES, of indigenous ethnicity, and who did not complete elementary school had a higher EC from whole grains compared to their respective counterparts. Red meat and processed meat had a higher EC among individuals from urban areas, of high-SES and of non-indigenous ethnicity, compared to their respective counterparts. EC from fruits differed by SES, with lower intake among low-SES individuals, whereas EC from vegetables differed between regions, with the North region having the lowest. Manufactured SSBs had a higher EC among men, individuals from urban areas, living in the North region, of high-SES, of non-indigenous ethnicity, or who completed elementary and middle school, relative to their respective counterparts; whereas homemade SSBs had a higher EC among individuals living in rural areas, in the South region, and from low-SES individuals, compared to their counterparts.

Changes from 2012 to 2016

The changes in EC from 2012 to 2016 were small, with average intakes of food group remaining similar in the four-year span. EC from cereals, and milk and dairy had statistically significant decreases, while EC from meat and animal products had a statistically significant increase (figure 1).

Data are from the Encuesta Nacional de Salud y Nutrición de Medio Camino 2016 (Ensanut-MC 2016), and the Ensanut 2012; n=11 640

Values are predicted means ± 95%Cis obtained with - margins - command from Stata from models adjusted by age group, sex, residence area, geographical region, socioeconomic level, indigenous ethnicity and weekend days

* p-values <0.05 comparing food group percent energy contribution of 2012 vs. 2016

SSBs: sugar-sweetened beverages; HSFAS: high in saturated fat and/or added sugar

Figure 1 Percent energy contribution from major food groups to total daily energy intake by survey year. Mexico, Ensanut 2012 and Ensanut-MC 2016 

Conclusion

In summary, the present analysis confirms the main conclusions from Ensanut of 2012, which draw attention to important deviations from the recommendations in regard to food group intake by the general Mexican population. Although subgroup analyses revealed among certain populations dietary patterns that are still heavily influenced by the traditional Mexican culture, all subgroups were adversely affected by a high intake of discretionary foods. And even though there are several policies in place aiming to improve the dietary quality among the population, there is still substantial room for improvement.