To assess the prevalence of promotions on foods and non-alcoholic drinks purchased by New Zealand households and to determine if they vary according to healthiness of products.

We undertook a cross-sectional analysis of Nielsen New Zealand Homescan® 2018/19 panel data. We conducted multivariate analyses to examine the variability in quantities of healthy v. unhealthy food and beverage products purchased on promotion. Promotion was self-reported by the panellist. Healthiness of products was measured by the Health Star Rating (HSR) system. We also carried out a subgroup analysis for beverages according to the threshold of < 5 g v. ≥ 5 g sugar per 100 ml content of products.

The Nielsen New Zealand Homescan® data were linked with two New Zealand Food Composition Databases (Nutritrack and the FOODfiles).

Food and beverage purchases data by 1800 panel households were used.

Overall, 46 % (1 803 601/3 940 458) of all purchases made were on promotion. Compared with purchases of food and beverage products with HSR < 3·5 (unhealthy), food and beverage products with HSR ≥ 3·5 (healthy) were significantly less likely to be on promotion (OR = 0·78, 95 % CI 0·77, 0·79). The subgroup analysis for beverages shows that products with < 5 g sugar per 100 ml were significantly less likely to be on promotion than those with ≥ 5 g sugar per 100 ml (OR = 0·77, 95 % CI 0·75, 0·79).

Policies to improve healthy food retailing should focus on increasing the promotion of healthier food and drink options in stores and supermarkets.

We determined the prevalence and identified predictors of food insecurity in four African countries.

Cross-sectional analyses at study enrolment.

From January 2013 to March 2020, people living with HIV (PLWH) and without HIV were enrolled at twelve clinics in Kenya, Uganda, Tanzania and Nigeria.

Participants reporting not having enough food to eat over the past 12 months or receiving <3 meals/d were defined as food insecure. Robust Poisson regression models were used to estimate unadjusted and adjusted prevalence ratios (aPR) and 95 % CI for predictors of food insecurity among all participants and separately among PLWH.

1694/3496 participants (48·5 %) reported food insecurity at enrolment, with no difference by HIV status. Food insecurity was more common among older participants (50+ v. 18–24 years aPR 1·35, 95 % CI 1·15, 1·59). Having 2–5 (aPR 1·14, 95 % CI 1·01, 1·30) or >5 dependents (aPR 1·17, 95 % CI 1·02, 1·35), and residing in Kisumu West, Kenya (aPR 1·63, 95 % CI 1·42, 1·87) or Nigeria (aPR 1·20, 95 % CI 1·01, 1·41) was associated with food insecurity. Residing in Tanzania (aPR 0·65, 95 % CI 0·53, 0·80) and increasing education (secondary/above education v. none/some primary education aPR 0·73, 95 % CI 0·66, 0·81) was protective against food insecurity. Antiretroviral therapy (ART)-experienced PLWH were more likely to be food secure irrespective of viral load.

Food insecurity was highly prevalent in our cohort though not significantly associated with HIV. Policies aimed at promoting education, elderly care, ART access in PLWH and financial independence could potentially improve food security in Africa.

To examine whether usual beverage intake was associated with sleep timing, duration and fragmentation among adolescents.

Usual beverage intake was assessed with a FFQ. Outcomes included sleep duration, midpoint (median of bed and wake times) and fragmentation, assessed with 7-d actigraphy. Sex-stratified linear regression was conducted with sleep characteristics as separate outcomes and quantiles of energy-adjusted beverage intake as exposures, accounting for age, maternal education, physical activity and smoking.

Mexico City.

528 adolescents residing in Mexico City enrolled in a longitudinal cohort.

The mean age (sd) was 14·4 (2·1) years; 48 % were male. Among males, milk and water consumption were associated with longer weekday sleep duration (25 (95 % CI 1, 48) and 26 (95 % CI 4, 47) more minutes, in the 4th compared to the 1st quartile); and higher 100 % fruit juice consumption was related to earlier weekday sleep timing (−22 (95 % CI −28, 1) minutes in the 1st compared to the last quantile; P = 0·03). Among females, soda was associated with higher sleep fragmentation (1·6 (95 % CI 0·4, 2·8) % in the 4th compared to the 1st), and coffee/tea consumption was related to shorter weekend sleep duration (−23 (95 % CI −44, 2) minutes in the 4th compared to the 1st).

Among females, adverse associations with sleep were observed for caffeinated drinks, while males with higher consumption of healthier beverage options (water, milk and 100 % juice) had evidence of longer and earlier-timed sleep. Potential mechanisms involving melatonin and tryptophan should be further investigated.

To examine how socio-demographic characteristics and diet quality vary with consumption of ultra-processed foods (UPF) in a cross-sectional nationally representative survey of Australian adults.

Using a 24-h recall, this cross-sectional analysis of dietary and socio-demographic data classified food items using the NOVA system, estimated the percentage of total energy contributed by UPFs and assessed diet quality using the Dietary Guideline Index (DGI–2013 total and components). Linear regression models examined associations between socio-demographic characteristics and diet quality with percentage of energy from UPF.

Australian Health Survey 2011–2013.

Australian adults aged ≥ 19 years (n 8209).

Consumption of UPF was higher among younger adults (aged 19–30 years), adults born in Australia, those experiencing greatest area-level disadvantage, lower levels of education and the second lowest household income quintile. No significant association was found for sex or rurality. A higher percentage of energy from UPF was inversely associated with diet quality and with lower DGI scores related to the variety of nutritious foods, fruits, vegetables, total cereals, meat and poultry, fish, eggs, nuts and seeds, legumes/beans, water and limits on discretionary foods, saturated fat and added sugar.

This research adds to the evidence on dietary inequalities across Australia and how UPF are detrimental to diet quality. The findings can be used to inform interventions to reduce UPF consumption and improve diet quality.

To examine associations between geographic information systems (GIS)-assessed accessibility to small food stores, shopping patterns and dietary behaviours among small food store customers.

Residential addresses and customer shopping patterns (frequency of shopping, and previous purchase of fruits and vegetables) were gathered through customer intercept surveys. Addresses were geocoded, and GIS-assessed distance and driving time from the participants’ residence to the store were calculated. Dietary status and behaviours were assessed using an objective non-invasive measure of skin carotenoids, the National Cancer Institute Fruit and Vegetable Screener, and items to assess sugary beverage intake. Associations between distance and driving time, demographics, shopping frequency, prior reported purchase of fruits and vegetables at the store and dietary behaviours were examined.

Small food stores (n 22) across North Carolina.

Cross-sectional convenience samples of English-speaking customers aged 18 years or older (n 692).

Participants living closer to the small store had lower income and formal education, were more likely to be Black, more likely to have previously bought fruits and vegetables at the store and more frequently shopped at the store. In adjusted models, skin carotenoids (n 644) were positively associated with distance to the store from home in miles (P = 0·01).

Customers who lived closer to the stores were more frequent shoppers and more likely to have previously purchased fruits and vegetables at the store yet had lower skin carotenoids. These results support continued efforts to examine how to increase the availability and promotion of healthful foods at small food retail stores.

This study aimed to verify the association between socio-economic and demographic characteristics and dietary patterns (DP) of children assisted by the Conditional Cash Transfer Program, Bolsa Família Program (BFP).

This is a cross-sectional study. DP were defined using a principal component analysis. The association of the predictive variables and DP was modelled using multilevel linear regression analysis.

This study was conducted in six municipalities from the State of Alagoas, Brazil.

The participants were children aged 6–24 months who were assisted by the BFP.

A total of 1604 children were evaluated. Four DP were identified (DP1, DP2, DP3 and DP4). DP1 is composed of traditional Brazilian food. DP2 is formed mostly from ultra-processed foods (UPF). DP3 consists of milk (non-breast) with added sugar, while DP4 consists of fresh and minimally processed foods. Caregivers with higher age and education (β = −0·008; (95 % CI −0·017, −0·000); β = −0·037; (95 % CI −0·056, −0·018), respectively) were negatively associated with DP2. We observed a negative association between households with food insecurity (β = −0·204; (95 % CI −0·331, −0·078)) and DP4 and a positive association between caregivers with higher age and education (β = 0·011; (95 % CI (0·003; 0·019); β = 0·043; (95 % CI 0·025, 0·061), respectively) and DP4.

This study identified the association between socio-economic inequities and DP early in life, with an early introduction of UPF, in children assisted by BFP in the State of Alagoas.

This study aimed to examine the impact of different dietary patterns on stroke outcomes among type 2 diabetes mellitus (T2DM) patients in China.

Participants were enrolled by a stratified random cluster sampling method in the study. After collecting dietary data using a quantified FFQ, latent class analysis was used to identify dietary patterns, and propensity score matching was used to reduce confounding effects between different dietary patterns. Binary logistic regression and conditional logistic regression were used to analyse the relationship between dietary patterns and stroke in patients with T2DM.

A cross-sectional survey available from December 2013 to January 2014.

A total of 13 731 Chinese residents aged 18 years or over.

Two dietary patterns were identified: 61·2 % of T2DM patients were categorised in the high-fat dietary pattern while 38·8 % of patients were characterised by the balanced dietary pattern. Compared with the high-fat dietary pattern, the balanced dietary pattern was associated with reduced stroke risk (OR = 0·63, 95 %CI 0·52, 0·76, P < 0·001) after adjusting for confounding factors. The protective effect of the balanced model did not differ significantly (interaction P > 0·05).

This study provides sufficient evidence to support the dietary intervention strategies to prevent stroke effectively. Maintaining a balanced dietary pattern, especially with moderate consumption of foods rich in quality protein and fresh vegetables in T2DM patients, might decrease the risk of stroke in China.

Characterising meat consumption in Switzerland across socio-demographic, lifestyle and anthropometric groups.

Representative national data from the menuCH survey (two 24-hour dietary recalls, anthropometric measurements and a lifestyle questionnaire) were used to analyse the total average daily intake of meat and main meat categories. Energy-standardised average intake (g/1000 kcal) was calculated and its association with 12 socio-demographic, lifestyle and anthropometric variables was investigated using multivariable linear regression.

Switzerland.

Totally, 2057 participants aged 18–75 years.

Average total meat intake was 109 g/d, which included 43 g/d of processed meat, 37 g/d of red meat and 27 g/d of white meat. Energy-standardised meat intake was highest for men, the Italian-language region and the youngest age group (18–29 years). Regression results showed significantly lower total meat and red meat consumption (g/1000 kcal) for women than men. However, there were no sex-specific differences for white meat. Total meat and white meat consumption were positively associated with the 18–29 age group, compared with 30–44 years, non-Swiss compared with Swiss participants and one-parent families with children compared with couples without children. Consumption of all categories of meat showed positive associations for BMI > 25 kg/m2 compared with BMI 18·5–25 kg/m2 and for French- and Italian-language regions compared with German-language region.

The current study reveals that there are significant differences in the amounts and types of meat consumed in Switzerland, suggesting that evidence-based risks and benefits of these categories need to be emphasised more in meat consumption recommendations.

Limiting SFA intake may minimise the risk of CHD. However, such reduction often leads to increased intake of carbohydrates. We aimed to evaluate associations and the interplay of carbohydrate and SFA intake on CHD risk.

Prospective cohort study.

We followed participants in the Hordaland Health Study, Norway from 1997–1999 through 2009. Information on carbohydrate and SFA intake was obtained from a FFQ and analysed as continuous and categorical (quartiles) variables. Multivariable Cox regression estimated hazard ratios (HR) and 95 % CI. Theoretical substitution analyses modelled the substitution of carbohydrates with other nutrients. CHD was defined as fatal or non-fatal CHD (ICD9 codes 410–414 and ICD10 codes I20–I25).

2995 men and women, aged 46–49 years.

Adjusting for age, sex, energy intake, physical activity and smoking, SFA was associated with lower risk (HRQ4 v. Q1 0·44, 95 % CI 0·26, 0·76, Ptrend = 0·002). For carbohydrates, the opposite pattern was observed (HRQ4 v. Q1 2·10, 95 % CI 1·22, 3·63, Ptrend = 0·003). SFA from cheese was associated with lower CHD risk (HRQ4 v. Q1 0·44, 95 % CI 0·24, 0·83, Ptrend = 0·006), while there were no associations between SFA from other food items and CHD. A 5 E% substitution of carbohydrates with total fat, but not SFA, was associated with lower CHD risk (HR 0·75, 95 % CI 0·62, 0·90).

Higher intake of predominantly high glycaemic carbohydrates and lower intake of SFA, specifically lower intake from cheese, were associated with higher CHD risk. Substituting carbohydrates with total fat, but not SFA, was associated with significantly lower risk of CHD.

To determine the association between ultra-processed food (UPF) intake and all-cause mortality in a representative sample of Spanish population.

Prospective cohort design in which follow-up lasted from baseline (1991) to mortality date or 31 December 2017, whichever was first. Dietary information was collected using a validated frequency questionnaire and categorised following the NOVA classification according to the extent of food processing. The association between consumption of UPF and mortality was analysed using Cox models. Isoenergetic substitution models were constructed to compare the health effects of the NOVA groups.

Cohort from the Diet and Risk of Cardiovascular Diseases (CVD) in Spain (DRECE) study, representative of the Spanish population

Totally, 4679 subjects between 5 and 59 years old.

Average consumption of UPF was 370·8 g/d (24·4 % of energy intake). After a median follow-up of 27 years, 450 deaths occurred. Those who consumed the highest amount of UPF had higher risk of mortality. For every 10 % of the energy intake from UPF consumption, an increase of 15 % in the hazard of all-cause mortality was observed (HR 1·15; (95 % CI 1·03, 1·27); P-value = 0·012). Substitution of UPF with minimally processed foods was significantly associated with a decreased risk of mortality.

An increase in UPF consumption was associated with higher risk of all-cause mortality in a representative sample of the Spanish population. Moreover, the theoretical substitution of UPF with unprocessed or minimally processed foods leads to a decrease in mortality. These results support the need to promote diets based on unprocessed or minimally processed foods.

The US Dietary Guidelines for Americans recommends increased consumption of the dairy group to three daily servings for ages 9+ years to help achieve adequate intakes of prominent shortfall nutrients. Identifying affordable, consumer-acceptable foods to replace dairy’s shortfall nutrients is important especially for people who avoid dairy.

Linear programming identified food combinations to replace dairy’s protein and shortfall nutrients. We examined cost, energy and dietary implications of replacing dairy with food combinations optimised for lowest cost, fewest kJ or the smallest amount of food by weight.

National Health and Nutrition Examination Survey (2011–2014).

Nationally representative sample of US population; 2 years and older (n 15 830).

Phase 1 (only dairy foods excluded): when optimised for lowest cost or fewest kJ, all non-dairy food replacements required large amounts (2·5–10 cups) of bottled/tap water. Phase 2 (dairy and unreasonable non-dairy foods excluded (e.g. baby foods; tap/bottled water): when intake of non-dairy foods was constrained to <90th percentile of current intake, the lowest cost food combination replacements for dairy cost 0·5 times more and provide 5·7 times more energy; the lowest energy food combinations cost 5·9 times more, provide 2·5 times more energy and require twice the amount of food by weight; and food combinations providing the smallest amount of food by weight cost 3·5 times more and provide five times more energy than dairy.

Identifying affordable, consumer-acceptable foods that can replace dairy’s shortfall nutrients at both current and recommended dairy intakes remains a challenge.

To examine the effect of the intervention implemented in the ToyBox-study on changes observed in age- and sex-specific BMI percentile and investigate the role of perinatal factors, parental perceptions and characteristics on this change.

A multicomponent, kindergarten-based, family-involved intervention with a cluster-randomised design. A standardised protocol was used to measure children’s body weight and height. Information was also collected from parents/caregivers via the use of validated questionnaires. Linear mixed effect models with random intercept for country, socio-economic status and school were used.

Selected preschools within the provinces of Oost-Flanders and West-Flanders (Belgium), Varna (Bulgaria), Bavaria (Germany), Attica (Greece), Mazowieckie (Poland) and Zaragoza (Spain).

A sample of 6268 preschoolers aged 3·5–5·5 years (51·9 % boys).

There was no intervention effect on the change in children’s BMI percentile. However, parents’ underestimation of their children’s actual weight status, parental overweight and mothers’ pre-pregnancy overweight/obesity were found to be significantly and independently associated with increases in children’s BMI percentile in multivariate modelling.

As part of a wide public health initiative or as part of a counseling intervention programme, it is important to assist parents/caregivers to correctly perceive their own and their children’s weight status. Recognition of excessive weight by parents/caregivers can increase their readiness to change and as such facilitate higher adherence to favourable behavioural changes within the family.

To assess the association between household food insecurity (FI) and major depressive episodes (MDE) amid Covid-19 pandemic in Brazil.

Cross-sectional study carried out with data from four consecutive population-based studies.

The study was conducted between May and June 2020, in Bagé, a Brazilian southern city. Household FI was measured using the short-form version of the Brazilian Food Insecurity Scale. Utilising the Patient Health Questionnaire-9, we used two different approaches to define MDE: the cut-off point of ≥ 9 and the diagnostic criteria proposed by the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR). Association between FI and MDE was analysed using crude and adjusted Poisson regression models.

1550 adults (≥ 20 years old).

The prevalence of household FI was 29·4 % (95 % CI 25·0, 34·4). MDE prevalence varied from 4·4 % (95 % CI 3·1, 6·0), when we used the DSM-IV-TR criteria to define this condition, to 9·6 % (95 % CI 7·3, 12·5) of the sample, when we used the cut-off point of ≥ 9 as definition. The prevalence of MDE was more than two times higher in those individuals living with FI, independent of the criteria adopted to define the outcome. Adjustment for potential confounders did not change the association’s magnitude.

Household FI has been positively associated with MDE amid Covid-19 pandemic, independent of socio-demographic characteristics of participants. Actions are needed to warrant basic living conditions to avoid FI and hunger and its consequences for the Brazilian population, especially those consequences linked to mental health disorders.

To examine the relationship between a posteriori dietary patterns in early childhood and alcohol consumption in adolescence.

Data were obtained from the Avon Longitudinal Study of Parents and Children (ALSPAC) prospective cohort study. Dietary information was obtained using FFQ at the age of 3 and 7 years. The association between dietary patterns, derived using principal components analysis and the Alcohol Use Disorders Identification Test (AUDIT) scores (to assess harmful intake) and frequency of alcohol consumption at the age of 17 years were examined. Secondary analysis considered sugar intake as a percentage of total energy intake.

Women who gave birth between 1 April 1991 and 31 December 1992 in the Avon area in southwest England were eligible for the ALSPAC cohort study.

Totally, 14 541 pregnancies were enrolled in ALSPAC during its initial recruitment phase. For this analysis, complete data were available for between 3148 and 3520 participants.

Adherence to the ‘healthy’ dietary pattern at both 3 and 7 years of age was positively associated with consuming more than one alcoholic drink per week at 17 years of age, whilst adherence to the ‘traditional’ dietary pattern at both ages was protective of harmful alcohol intake at 17 years of age. Sugar intake was not associated with either alcohol outcome after adjustment for ethnicity, maternal level of education, parental social class and maternal AUDIT score.

For the population studied, changes to diet in early childhood are unlikely to have an impact on harmful alcohol use in adolescence given the lack of consistency across the results.

To assess the association between soybean consumption and anaemic status in Central Java, Indonesia.

As part of an overarching sanitation improvement intervention in Central Java, Indonesia, we conducted a cross-sectional study in four rural villages. The study consisted of a 24-h food recall, anthropometric measurements, blood Hb measurement and stool sampling to test for soil-transmitted helminth (STH) infection status. A binary logistic regression model was fitted to test the association between soybean consumption and anaemic status after adjusting for socio-demographic factors, STH infection, dietary diversity and anthropometric status.

This study took place in four rural villages of Wonosobo regency, Central Java, Indonesia.

Participants were rural villagers aged between 15 and 49 years.

A total sample size of 763 was attained, of which 231 were anaemic. The prevalence of anaemia was 30·2 % among men and women of reproductive age, and highest among young males. Consumption of soybean was high (79·8 %). After adjusting for covariates, the protective association between soybean consumption and anaemia was statistically significant (AOR = 0·53, 95 % CI = 0·30, 0·95, P < 0·05). There was a positive association with anaemia among underweight (AOR = 2·75, 95 % CI = 1·13, 6·69, P < 0·05) and those with high diet diversity (AOR = 1·40, 95 % CI = 1·00, 1·97, P < 0·05).

Our results were consistent with studies from other countries finding a protective association between soybean consumption and anaemia. This association appeared stronger for tofu than for tempeh. The prevalence of anaemia in rural Central Java is relatively consistent with nation-wide statistics indicating that interventions targeting anaemia are still largely required.

The present work was performed to analyse the association of dietary patterns with glycaemic control (Hb A1c < 7 %) in a large group of Chinese adults aged between 45 and 59 years.

Habitual dietary intakes in the preceding 12 months were assessed by well-trained interviewers using a validated semi-quantitative FFQ. Factor analysis with principal component method was used to obtain the dietary patterns, and the associations between dietary patterns and glycaemic control were determined using multivariable logistic regression models. Poor glycaemic control was defined as HbA1c ≥ 7·0.

Despite decades of research, data on the relationship between dietary patterns and glycaemic control (HbA1c < 7 %) in China are sparse.

A total of 1739 participants aged 45–59 years from Hangzhou were included in the final analysis.

Three dietary patterns were ascertained and labelled as traditional southern Chinese, Western and grains-vegetables patterns. After controlling of the possible confounders, participants in the highest quartile of Western pattern scores had greater OR for HbA1c ≥ 7·0 (OR = 1·05; (95 % CI 1·000, 1·095); P = 0·048) than did those in the lowest quartile. Compared with those in the lowest quartile of grains-vegetables pattern, participants in the highest quartile had lower OR for HbA1c ≥ 7·0 (OR = 0·82; (95 % CI 0·720, 0·949); P = 0·038). Besides, no significant relationship between the traditional southern Chinese pattern and HbA1c ≥ 7·0 was observed (P > 0·05).

This study indicated that the Western pattern was associated with a higher risk, and the grains-vegetables pattern was associated with a lower risk for HbA1c ≥ 7·0. Future prospective studies are needed to confirm our findings.

This study investigated associations between types and food sources of protein with overweight/obesity and underweight in Ethiopia.

We conducted a cross-sectional dietary survey using a non-quantitative FFQ. Linear regression models were used to assess associations between percentage energy intake from total, animal and plant protein and BMI. Logistic regression models were used to examine the associations of percentage energy intake from total, animal and plant protein and specific protein food sources with underweight and overweight/obesity.

Addis Ababa, Ethiopia.

1624 Ethiopian adults (992 women and 632 men) aged 18–49 years in selected households sampled using multi-stage random sampling from five sub-cities of Addis Ababa.

Of the surveyed adults, 31 % were overweight or obese. The majority of energy intake was from carbohydrate with only 3 % from animal protein. In multivariable-adjusted linear models, BMI was not associated with percentage energy from total, plant or animal protein. Total and animal protein intake were both associated with lower odds of overweight/obesity (OR per 1 % energy increment of total protein 0·92; 95 % CI: 0·86, 0·99; P = 0·02; OR per 1 % energy increment of animal protein 0·89; 95 % CI: 0·82, 0·96; P = 0·004) when substituted for carbohydrate and adjusted for socio-demographic covariates.

Increasing proportion of energy intake from total protein or animal protein in place of carbohydrate could be a strategy to address overweight and obesity in Addis Ababa; longitudinal studies are needed to further examine this potential association.

To explore long-term trends in height, weight and BMI across birth cohorts among Indian women aged 15–30 years.

Nationally representative cross-sectional surveys.

Data from three National Family Health Surveys were conducted in 1998–1999, 2005–2006 and 2015–2016. Height and weight were modelled jointly, employing a multivariate regression model with age and birth cohorts as explanatory variables. The largest birth cohort (born 1988–1992) was the reference cohort. Stratified analyses by place of residence and by marital status and dichotomised parity were also performed.

437 753 non-pregnant women aged 15–30 years.

The rate of increase in height, weight and BMI differed across birth cohorts. The rate of increase was much lower for height than weight, which was reflected in an increasing trend in BMI across all birth cohorts. In the stratified analyses, increase in height was found to be similar across urban and rural areas. Rural women born in the latest birth cohort (1998–2001) were lighter, whereas urban women were heavier compared to the reference cohort. A relatively larger increase in regression coefficients was observed among women born between 1978 and 1982 compared to women born between 1973 and 1977 when considering unmarried and nulliparous ever-married women and, one cohort later (1983–1987 v. 1978–1982), among parous ever-married women.

As the rate of increase was much larger for weight than for height, increasing trends in BMI were observed across the birth cohorts. Thus, cohort effects show an important contributory role in explaining increasing trends in BMI among young Indian women.

This study aimed to identify health behaviours that determine adolescent’s adherence to the Mediterranean diet (MD) through a decision tree statistical approach.

Cross-sectional study, with data collected through a self-fulfilment questionnaire with five sections: (1) eating habits; (2) adherence to the MD (KIDMED index); (3) physical activity; (4) health habits and (5) socio-demographic characteristics. Anthropometric and blood pressure data were collected by a trained research team. The Automatic Chi-square Interaction Detection (CHAID) method was used to identify health behaviours that contribute to a better adherence to the MD.

Eight public secondary schools, in Algarve, Portugal.

Adolescents with ages between 15 and 19 years (n 325).

According to the KIDMED index, we found a low adherence to MD in 9·0 % of the participants, an intermediate adherence in 45·5 % and a high adherence in 45·5 %. Participants that regularly have breakfast, eat vegetable soup, have a second piece of fruit/d, eat fresh or cooked vegetables 1 or more times a day, eat oleaginous fruits at least 2 to 3 times a week, and practice sports and leisure physical activities outside school show higher adherence to the MD (P < 0·001).

The daily intake of two pieces of fruit and vegetables proved to be a determinant health behaviour for high adherence to MD. Strategies to promote the intake of these foods among adolescents must be developed and implemented.