The main aim of the present study was to identify food consumption in Sri Lankan adults based on serving characteristics.

Cross-sectional study. Fruits, vegetables, starch, meat, pulses, dairy products and added sugars in the diet were assessed with portion sizes estimated using standard methods.

Twelve randomly selected clusters from the Sri Lanka Diabetes and Cardiovascular Study.

Six hundred non-institutionalized adults.

The daily intake of fruit (0·43), vegetable (1·73) and dairy (0·39) portions were well below national recommendations. Only 3·5 % of adults consumed the recommended 5 portions of fruits and vegetables/d; over a third of the population consumed no dairy products and fewer than 1 % of adults consumed 2 portions/d. In contrast, Sri Lankan adults consumed over 14 portions of starch and 3·5 portions of added sugars daily. Almost 70 % of those studied exceeded the upper limit of the recommendations for starch intake. The total daily number of meat and pulse portions was 2·78.

Dietary guidelines emphasize the importance of a balanced and varied diet; however, a substantial proportion of the Sri Lankan population studied failed to achieve such a recommendation. Nutrition-related diseases in the country may be closely correlated with unhealthy eating habits.

To develop a basis for building models that can examine the impact of organic food (OF) choices on maternal and offspring health, including identification of factors associated with OF consumption and underlying dietary patterns.

Dietary intake was collected for the preceding month from an FFQ in mid-pregnancy and information on sociodemographic characteristics was collected from telephone interviews during pregnancy. From a question about OF consumption in the FFQ, including six food categories, an OF preference index was calculated. Latent variables that captured the variability in OF choices in relation to dietary intake were defined.

The Danish National Birth Cohort (DNBC), 1996–2002.

Pregnant women from DNBC (n 60 773).

We found that frequent OF use was highly associated with age, occupational status, urbanization, smoking and vegetarianism. By principal components analysis we identified two eating patterns, a ‘Western dietary pattern’ and a ‘Prudent dietary pattern’, that explained 14·2 % of the variability in data. Frequent OF users consumed a more ‘prudent’ diet compared with non-users and had significantly higher intakes of vegetables (+67 %), fibre (+13 %) and n-3 fatty acids (+11 %) and less saturated fat (−8 %).

Frequent OF users seemed to have a healthier lifestyle than non-users. These findings highlight a major challenge in observational studies examining the impact of OF consumption on health due to potentially irremediable confounding factors.

Adequate fluid intake has been well documented as important for health but whether it has adverse effects on overall energy and sugar intakes remains under debate. Many dietary studies continue to refrain from reporting on beverage consumption, which the present study aimed to address.

A cross-sectional survey investigated self-reported measures of dietary intake and anthropometric measurements.

Primary and secondary schools in south-west London, UK.

Boys and girls (n 248) aged 9–13 years.

Boys consumed 10 % and girls consumed 9 % of their daily energy intake from beverages and most children had total sugar intakes greater than recommended. Beverages contributed between a quarter and a third of all sugars consumed, with boys aged 11–13 years consuming 32 % of their total sugar from beverages. There was a strong relationship between consumption of beverages and energy intake; however, there was no relationship between beverage type and either BMI or BMI Z-score. Fruit juices and smoothies were consumed most frequently by all girls and 9–10-year-old boys; boys aged 11–13 years preferred soft drinks and consumed more of their daily energy from soft drinks. Milk and plain water as beverages were less popular.

Although current health promotion campaigns in schools merit the attention being given to improving hydration and reducing soft drinks consumption, it may be also important to educate children on the energy and sugar contents of all beverages. These include soft drinks, as well as fruit juices and smoothies, which are both popular and consumed regularly.

To evaluate the validity and reproducibility of a 122-item interviewer-administered quantitative FFQ developed to determine food and nutrient intakes of adults in Botswana.

Relative validity of the FFQ was evaluated by comparing nutrient and food group intakes against four non-consecutive 24 h recalls administered over 12 months. The FFQ was repeated after 1 year to assess reproducibility.

Kanye, Botswana.

Seventy-nine adults aged 18–75 years.

Spearman correlation coefficients for the validity of energy-adjusted nutrients ranged from 0·42 (carbohydrate) to 0·49 (protein) for macronutrients and from 0·23 (Fe) to 0·44 (PUFA) for micronutrients. Exact agreement of quartile distribution for nutrients between the FFQ and recalls ranged from 27 % to 72 %. Weighted kappa values were lowest for retinol (0·13), Fe (0·22) and β-carotene (0·25) and ranged from 0·33 (SFA) to 0·59 (folate) for other nutrients (energy, carbohydrate, protein, fat, Ca and vitamin E). Spearman correlation coefficients between the recalls and FFQ for food groups ranged from 0·18 (dark green leafy and yellow vegetables) to 0·58 (poultry). Reproducibility correlation coefficients (energy-adjusted) varied between 0·39 for retinol and 0·66 for vitamin E, with most values falling between 0·50 and 0·60.

The FFQ had good relative validity for estimating habitual food group and nutrient intakes, but was poor for some micronutrients (Fe, retinol and β-carotene) and foods (fruits and dark green leafy vegetables).

The health benefits of exercise are clear. In targeting interventions it would be valuable to know whether characteristic patterns of physical activity (PA) are associated with particular population subgroups. The present study used cluster analysis to identify characteristic hourly PA patterns measured by accelerometer.

Cross-sectional design.

Objectively measured PA in Hong Kong adults.

Four-day accelerometer data were collected during 2009 to 2011 for 1714 participants in Hong Kong (mean age 44·2 years, 45·9 % male).

Two clusters were identified, one more active than the other. The ‘active cluster’ (n 480) was characterized by a routine PA pattern on weekdays and a more active and varied pattern on weekends; the other, the ‘less active cluster’ (n 1234), by a consistently low PA pattern on both weekdays and weekends with little variation from day to day. Demographic, lifestyle, PA level and health characteristics of the two clusters were compared. They differed in age, sex, smoking, income and level of PA required at work. The odds of having any chronic health conditions was lower for the active group (adjusted OR = 0·62, 95 % CI 0·46, 0·84) but the two groups did not differ in terms of specific chronic health conditions or obesity.

Implications are drawn for targeting exercise promotion programmes at the population level.

To develop a feasible, valid, reliable web-based instrument to objectively evaluate school meal quality in Swedish primary schools.

The construct ‘school meal quality’ was operationalized by an expert panel into six domains, one of which was nutritional quality. An instrument was drafted and pilot-tested. Face validity was evaluated by the panel. Feasibility was established via a large national study. Food-based criteria to predict the nutritional adequacy of school meals in terms of fat quality, iron, vitamin D and fibre content were developed. Predictive validity was evaluated by comparing the nutritional adequacy of school menus based on these criteria with the results from a nutritional analysis. Inter-rater reliability was also assessed.

The instrument was developed between 2010 and 2012. It is designed for use in all primary schools by school catering and/or management representatives.

A pilot-test of eighty schools in Stockholm (autumn 2010) and a further test of feasibility in 191 schools nationally (spring 2011).

The four nutrient-specific food-based criteria predicted nutritional adequacy with sensitivity ranging from 0·85 to 1·0, specificity from 0·45 to 1·0 and accuracy from 0·67 to 1·0. The sample in the national study was statistically representative and the majority of users rated the questionnaire positively, suggesting the instrument is feasible. The inter-rater reliability was fair to almost perfect for continuous variables and agreement was ≥67 % for categorical variables.

An innovative web-based system to comprehensively monitor school meal quality across several domains, with validated questions in the nutritional domain, is available in Sweden for the first time.

To investigate the nutritional impact of a community-based programme that focused on social cohesion and action.

The change in nutritional status of children aged 12–60 months was examined over a period of 3 years in Makueni District in Eastern Province of Kenya in six communities in which an intervention programme of Participatory Learning and Action was introduced and in ten communities in which only basic preparations were made but no intervention was started.

The intervention was part of the Government of Kenya Community Based Nutrition Programme and was supported by the Government of Denmark.

Children aged 12–60 months.

Among communities without intervention there were similar levels of underweight (mean Z-score: −1·63 v. −1·50 (NS); % with Z-score < –2: 36·6 % v. 34·5 % (NS)) and stunting (mean Z-score: −2·0 v. −1·99 (NS); % with Z-score < –2: 44·3 % and 47·4 % (NS)) at baseline and after 3 years. By contrast, among communities who had received interventions, there were significant improvements after 3 years in the levels of underweight (mean Z-score: −1·66 v. −1·37 (P < 0·02); % with Z-score < –2: 42·9 % v. 31·4 % (P < 0·035)) and stunting (mean Z-score: −2·05 v. −1·59 (P < 0·05); % with Z-score < –2: 52·7 % v. 39·7 % (P < 0·02)).

The results indicate considerable potential for using Participatory Learning and Action as a community-based approach to effectively address child undernutrition. It is suggested that these interventions are developed, implemented and evaluated more widely as a mean of tackling childhood undernutrition and improving child survival and development.

To investigate which anthropometric measure of overweight status, BMI or waist-to-height ratio (WHtR), is most closely associated with parents’ perception of their child's overweight status.

The sensitivity and specificity of parental perception against child-specific BMI and WHtR definitions of overweight were tested.

Primary schools in Queensland, Australia.

Boys and girls aged 9·00–11·99 years (n 1431).

Of the 138 boys and 202 girls who were classified as overweight according to BMI, only 27·5 % (boys) and 22·7 % (girls) were also perceived as overweight by their parents. Using WHtR, 206 boys and 333 girls were classified as overweight, of whom only 21·9 % and 13·8 %, respectively, were perceived as overweight. Perception of overweight was underestimated in approximately 15 % of boys and 21 % of girls when compared with BMI. Underestimation was higher when compared with WHtR: 25 % (boys) and 39 % (girls). Overweight prevalence was significantly lower according to perception than according to BMI or WHtR. Mother's education level was significantly associated with accurate perception of overweight status (P < 0·001).

The sensitivity of parental perception of child overweight was higher when BMI was used. However, emphasis needs to be placed on using WHtR as an actual measure of overweight because high central adiposity is associated with increased risk of CVD. The combined use of WHtR, body-shape images rather than word responses regarding perception and public health messages that educate parents and children about body shape and associated health risks may be the best combination in improving parents’ perception of their child's overweight status.

Prevalence of acute malnutrition is classically estimated by the proportion of children meeting a case definition in a representative population sample. In 1995 the WHO proposed the PROBIT method, based on converting parameters of a normally distributed variable to cumulative probability, as an alternative method requiring a smaller sample size. The present study compares classical and PROBIT methods for estimating the prevalence of global, moderate and severe acute malnutrition (GAM, MAM and SAM) defined by weight-for-height Z-score (WHZ) or mid-upper arm circumference (MUAC).

Bias and precision of classical and PROBIT methods were compared by simulating a total of 1·26 million surveys generated from 560 nutrition surveys.

Data used for simulation were derived from nutritional surveys of children aged 6–59 months carried out in thirty-one countries around the world.

Data of 459 036 children aged 6–59 months from representative samples were used to generate simulated populations.

The PROBIT method provided an estimate of GAM, MAM and SAM using WHZ or MUAC proportional to the true prevalence with a small systematic overestimation. The PROBIT method was more precise than the classical method for estimating the prevalence for GAM, MAM and SAM by WHZ or MUAC for small sample sizes (i.e. n<150 for SAM and GAM; n<300 for MAM), but lost this advantage when sample sizes increased.

The classical method is preferred for estimating acute malnutrition prevalence from large sample surveys. The PROBIT method may be useful in sentinel-site surveillance systems with small sample sizes.

To determine the main dietary patterns of pregnant women during each of the three trimesters of pregnancy and to examine associated nutrient intakes.

Participants completed a 3 d food diary during each trimester of pregnancy. Thirty-six food groups were created and dietary patterns were derived using k-means cluster analysis.

National Maternity Hospital, Dublin, Ireland.

Two hundred and eighty-five healthy pregnant women aged between 20 and 41 years.

Two dietary patterns were identified at each time point. They were labelled ‘Unhealthy’ (n 143, 150 and 155 at trimester 1, 2 and 3, respectively) and ‘Health Conscious’ (n 142, 135 and 130 at trimester 1, 2 and 3, respectively). Women in the ‘Health Conscious’ cluster were significantly older, had lower BMI and were higher educated than those in the ‘Unhealthy’ cluster. Of those in the ‘Unhealthy’ cluster in the first trimester (n 143), 103 (72·0 %) continued in this dietary pattern into trimester 2 and eighty-one (56·6 %) continued into trimester 3. Of those in the ‘Health Conscious’ cluster in trimester 1 (n 142), ninety-five (66·9 %) continued in this dietary pattern into trimester 2 and sixty-nine (48·6 %) continued into trimester 3.

Cluster analysis produced two clearly defined dietary patterns at each stage of pregnancy. Knowledge of maternal dietary patterns is important for the development of pregnancy-specific dietary guidelines. Identifying women with an ‘Unhealthy’ dietary pattern in early pregnancy affords the opportunity for a dietary intervention which may positively impact both maternal and infant health.

The present study investigated Guarani village interviewees’ diet changes over time, their perceptions about the changes and the effects of these changes on their health.

The study employed qualitative methods with a sample of Guarani Indians selected by snowball sampling. Ethnographic methods and techniques included field diaries, informal and unstructured interviews and participant observation.

The Tenondé Porã Guarani village is located in the district of Parelheiros, São Paulo, Brazil. Interviews were conducted from July 2008 to December 2009.

Fifteen Guarani Indians, males and females in age categories ranging from youths to elders, took part in the study.

The interviewees reported changes in how food was obtained, the occurrence of food substitutions and food species abandonment, recipe changes and the introduction of new foods. Some ritual use of plants was maintained. Disease frequency was found to increase because of this change and a lack of obedience to Nhanderu (the Guarani God). A lack of space for daily traditional activities (e.g. farming, hunting) was found to result in sedentary lifestyles.

The village location was a key factor in the Guarani diet change, although some rituals related to available plants were preserved.

To explore mediators of gender and educational differences in sugar-sweetened soft drinks consumption (SDC) and whether gender and level of future education moderate the associations of accessibility, modelling, attitudes and preferences with SDC.

A cross-sectional school-based survey within the Fruits and Vegetables Makes the Marks (FVMM) project from 2005.

The questionnaires were completed by the pupils in the classroom guided by a trained project worker during one class session. The questionnaire included questions on SDC (times/week), the potential mediators and moderators. Multilevel linear regression models were used to calculate the mediating and moderating effects.

A total of 2870 children in 9th and 10th grade (mean age 15·5 years) at thirty-three Norwegian secondary schools were included in the present study.

Girls (B = −1·06) and pupils planning higher education (B = −0·69) reported lower frequency of SDC. The strongest mediators were accessibility and modelling for future educational plans differences (explaining alone respectively 69 % and 44 %) and attitudes and preferences for gender differences (explaining alone respectively 57 % and 51 %). Significant but small moderating effects were found, and all associations between the mediators and SDC were in the same direction for both genders and for those with and without plans of higher future education.

Preferences and modelling may contribute to gender and educational differences in SDC. The small moderating effects indicate that interventions aiming to reduce SDC can target the same mediators for boys and girls and children planning different levels of future education.

To compare food choices and nutrient intakes of pupils taking a school lunch or a packed lunch in eighty secondary schools in England, following the introduction of the food-based and nutrient-based standards for school food.

Cross-sectional data collected between October 2010 and April 2011. Pupils’ lunchtime food choices were recorded over five consecutive days.

Secondary schools, England.

A random selection of 5925 pupils having school lunches and 1805 pupils having a packed lunch in a nationally representative sample of eighty secondary schools in England.

The differences in the specific types of food and drink consumed by the two groups of pupils are typical of differences between a hot and cold meal. On average, school lunches as eaten contained significantly more energy, carbohydrate, protein, fibre, vitamin A, folate, Fe and Zn than packed lunches, and 8 % less Na.

Although neither school lunches nor packed lunches provided the balance of nutrients required to meet the nutrient-based standards (based on about one-third of daily energy and nutrient requirements), school lunches generally had a healthier nutrient profile, with lower Na and percentage of energy from fat, and higher fibre and micronutrient content. These differences were greater than those reported prior to the introduction of compulsory standards for school lunches. In order to ensure more pupils have a healthy lunch, schools could introduce and enforce a packed lunch policy or make school meals the only option at lunchtime.

It is likely that there are substantial subconscious effects of organizations’ efforts to associate their products with sport via sponsorships, but most research methods are unable to capture these effects. The present study employed a novel projective technique to explore children's implicit associations between popular sports and a range of sports sponsors.

Children participated in an activity using magnets bearing the logos of numerous sports and sponsors. They were invited to arrange the magnets on a whiteboard without being advised that the activity related to sponsorship.

Perth, Western Australia.

Children (n 164) aged 5–12 years.

Three-quarters (76 %) of the children aligned at least one correct sponsor magnet with the relevant sport. Just over half the children (54 %) correctly matched the most popular sport (an Australian Football League team) with its relevant sponsor (a fast-food chain).

Given the unstructured nature of the projective task, the results provide some support for the argument that sports sponsorship can effectively reach child audiences. This is of concern given the current extent of sponsorship by alcohol and fast-food companies.

Exclusive breast-feeding is estimated to reduce infant mortality in low-income countries by up to 13 %. The aim of the present study was to determine the risk factors associated with suboptimal breast-feeding practices in Pakistan.

A cross-sectional study using data extracted from the multistage cluster sample survey of the Pakistan Demographic and Health Survey 2006–2007.

A nationally representative sample of households.

Last-born alive children aged 0–23 months (total weighted sample size 3103).

The prevalences of timely initiation of breast-feeding, bottle-feeding in children aged 0–23 months, exclusive breast-feeding and predominant breast-feeding in infants aged 0–5 months were 27·3 %, 32·1 %, 37·1 % and 18·7 %, respectively. Multivariate analysis indicated that working mothers (OR = 1·48, 95 % CI 1·16, 1·87; P = 0·001) and mothers who delivered by Caesarean section (OR = 1·95, 95 % CI 1·30, 2·90; P = 0·001) had significantly higher odds for no timely initiation of breast-feeding. Mothers from North West Frontier Province were significantly less likely (OR = 0·37, 95 % CI 0·23, 0·59; P < 0·001) not to breast-feed their babies exclusively. Mothers delivered by traditional birth attendants had significantly higher odds to predominantly breast-feed their babies (OR = 1·96, 95 % CI 1·18, 3·24; P = 0·009). The odds of being bottle-fed was significantly higher in infants whose mothers had four or more antenatal clinic visits (OR = 1·93, 95 % CI 1·46, 2·55; P < 0·001) and belonged to the richest wealth quintile (OR = 2·41, 95 % CI 1·62, 3·58; P < 0·001).

The majority of Pakistani mothers have suboptimal breast-feeding practices. To gain the full benefits of breast-feeding for child health and nutrition, there is an urgent need to develop interventions to improve the rates of exclusive breast-feeding.

It is not clear whether measuring waist circumference in clinical practice is problematic because the measurement error is unclear, as well as what constitutes a clinically relevant change. The present study aimed to summarize what is known from state-of-the-art research.

To identify the magnitude of the measurement error of waist circumference measurements from the literature, a search was conducted in PubMed from 1975 to February 2011.

The measurement error may vary between 0·7 cm and 15 cm. Taking a realistic range of measurable waist circumference into account (60–135 cm), we argue that a short-term clinically relevant change in waist circumference of 5 % may lie between 3·0 and 6·8 cm and a maintained clinically relevant change of 3 % between 1·8 and 4·1 cm.

Based on these results, we conclude it may be difficult to distinguish clinically relevant change from measurement error in individual subjects, due to the large measurement error and unclear definition of clinically relevant change. More research is needed to address these gaps in knowledge. To minimize measurement error, we recommend using a uniform measurement protocol, training and repeated measurements.

Few studies have investigated the association between adolescents’ socio-economic status (SES) and health behaviours in developing countries. Socio-economic differences in breakfast eating, fruit and vegetable consumption and physical activity were investigated among Ghanaian adolescents.

A school-based survey of 12–18-year-olds (n 1195, response rate 89·7 %) was conducted in 2008. Logistic regression analyses were applied to study the associations between breakfast eating, fruit and vegetable consumption and physical activity, and SES.

Southern Ghana, West Africa.

School pupils aged 12–18 years.

Of all adolescents, 31 % took breakfast less than 4 d/week, over half (56 %) and 48 %, respectively, rarely ate fruits and vegetables. Younger adolescents (12–15-year-olds) consumed fruits and vegetables frequently compared with older ones (16–18-year-olds). Boys were more likely to participate in physical activity than girls. The probability of frequent breakfast eating was higher in adolescents from more affluent backgrounds than in those from less affluent ones. Compared with those whose mothers were illiterate, both tertiary and primary maternal educational attainment increased the probability of frequent fruit intake. Similarly, tertiary educational attainment increased the likelihood of frequent vegetable intake. Compared with those with unemployed mothers, mothers’ low employment grade was related to higher probability of frequent fruit intake. High school performance was associated with frequent fruit intake, whereas high or medium school performance increased the likelihood of vegetable intake compared with low school performance.

Interventions are needed to improve breakfast intake, fruit and vegetable consumption and physical activity in order to prevent degenerative diseases among Ghanaian adolescents.

To assess the relationship between parental education level and the consumption frequency of obesity-related foods in European children.

The analysis was based on data from the cross-sectional baseline survey of a prospective cohort study. The effects of parental education on food consumption were explored using analysis of covariance and logistic regression.

Primary schools and pre-schools of selected regions in Italy, Estonia, Cyprus, Belgium, Sweden, Hungary, Germany and Spain.

Participants (n 14 426) of the IDEFICS baseline cohort study aged 2 to 9 years.

Parental education level affected the intake of obesity-related foods in children. Children in the low and medium parental education level groups had lower odds of more frequently eating low-sugar and low-fat foods (vegetables, fruits, pasta/noodles/rice and wholemeal bread) and higher odds of more frequently eating high-sugar and high-fat foods (fried potatoes, fruits with sugar and nuts, snacks/desserts and sugared beverages; P < 0·001). The largest odds ratio differences were found in the low category (reference category: high) for vegetables (OR = 0·56; 95 % CI 0·47, 0·65), fruits (OR = 0·56; 95 % CI 0·48, 0·65), fruits with sugar and nuts (OR = 2·23; 95 % CI 1·92, 2·59) and sugared beverages (OR = 2·01; 95 % CI 1·77, 2·37).

Low parental education level was associated with intakes of sugar-rich and fatty foods among children, while high parental education level was associated with intakes of low-sugar and low-fat foods. These findings should be taken into account in public health interventions, with more targeted policies aiming at an improvement of children's diet.

A variety of methods are available for defining undernutrition (thinness/underweight/under-fat) and overnutrition (overweight/obesity/over-fat). The extent to which these definitions agree is unclear. The present cross-sectional study aimed to assess agreement between widely used methods of assessing nutritional status in children and adolescents, and to examine the benefit of body composition estimates.

The main objective of the cross-sectional study was to assess underweight, overweight and obesity using four methods: (i) BMI-for-age using WHO (2007) reference data; (ii) BMI-for-age using Cole et al. and International Obesity Taskforce cut-offs; (iii) weight-for-age using the National Centre for Health Statistics/WHO growth reference 1977; and (iv) body fat percentage estimated by bio-impedance (body fat reference curves for children of McCarthy et al., 2006). Comparisons were made between methods using weighted kappa analyses.

Rural South Africa.

Individuals (n 1519) in three age groups (school grade 1, mean age 7 years; grade 5, mean age 11 years; grade 9, mean age 15 years).

In boys, prevalence of unhealthy weight status (both under- and overnutrition) was much higher at all ages with body fatness measures than with simple anthropometric proxies for body fatness; agreement between fatness and weight-based measures was fair or slight using Landis and Koch categories. In girls, prevalence of unhealthy weight status was also higher with body fatness than with proxies, although agreement between measures ranged from fair to substantial.

Methods for defining under- and overnutrition should not be considered equivalent. Weight-based measures provide highly conservative estimates of unhealthy weight status, possibly more conservative in boys. Simple body composition measures may be more informative than anthropometry for nutritional surveillance of children and adolescents.

To assess the validity of a 161-item quantitative FFQ (QFFQ) that was developed to evaluate dietary risk factors for a colorectal adenoma case–control study.

A cross-sectional validation study of the QFFQ against 4 d food diary using Pearson correlation coefficients, cross-classification, weighted κ statistics and Bland–Altman plotting.

Two hospitals in São Paulo, Brazil.

Ninety-seven healthy Japanese-Brazilian adults (40–75 years) were recruited. One participant was excluded from the analysis due to unusual energy intake report.

Mean daily nutrient intakes from the QFFQ were higher than from the food diary. The mean Pearson correlation coefficient for nutrient intakes between the QFFQ and the average of the 4 d food diary was 0·43, and increased to 0·45 after correcting correlations for attenuation due to residual day-to-day variation in the food diary measurements. Adjustment for total energy and further adjustment for age and gender decreased the correlation; however, 77 % of observations remained in the same or adjacent quartiles with a mean weighted κ of 0·22. Bland–Altman plots on loge-transformed data showed no linear trend between the differences and means for energy, fat, protein, total folate and vitamin C. Compared with the food diary, the QFFQ showed consistently reasonable performance for dietary fibre, total folate, retinol, riboflavin and vitamin C.

This investigation supports the relative validity of the QFFQ as a method for assessing long-term dietary intake. The instrument will be a useful tool in the analysis of diet–adenoma associations in the case–control study.