To develop a household-level diet quality indicator (HDQI) using the Salvadorian dietary guidelines to assess the dietary quality of households in vulnerable communities in El Salvador.

The Salvadorian dietary guidelines were reviewed and eighteen HDQI components were identified (nine foods and nine nutrients). The components were evaluated using a proportional scoring system from 0 to 1, penalizing over- and under-consumption, where appropriate. The HDQI was validated in consultations with experts in El Salvador and by statistical analyses of the study sample data. Dietary variety and energy, nutrient and food intakes were compared among households above and below the median HDQI score using Student's t test.

Vulnerable, border communities in El Salvador.

Households (n 140) provided food consumption information using an FFQ and sociodemographic data.

The mean HDQI score was 63·5, ranging from 43·6 to 90·0. The indicator showed a positive, significant association with the dietary variety components. The statistical associations of the indicator with the energy and nutrient components were as expected.

Based on the indicator's demonstrated face validity and the results of the expert consultations, the indicator is suggested as a good measure of diet quality for households in El Salvador.

To determine if dietary, supplemental and total vitamin D intakes in the USA are influenced by income, race/ethnicity or gender.

Cross-sectional. US vitamin D intakes were estimated by poverty income ratio (PIR), race/ethnicity and gender using 24 h dietary intake recalls and dietary supplement use questionnaires. Statistical analyses of weighted data were performed using SAS (version 9·2) to estimate means and their standard errors. Race and ethnic intake differences controlling for PIR, gender and age were assessed by ANCOVA.

Adults aged ≥19 years.

The 2007–2010 National Health and Nutrition Examination Survey, USA.

Total (dietary and supplement) vitamin D intake was greater in the high (10·0 (se 0·30) μg/d) v. the medium (7·9 (se 0·3) μg/d) or the low (8·0 (se 0·3) μg/d) PIR categories. Total vitamin D intake of non-Hispanic Whites (10·6 (se 0·4) μg/d) was greater than that of Hispanics (8·1 (se 0·3) μg/d) and non-Hispanic Blacks (7·1 (se 0·3) μg/d). Supplemental vitamin D intake was greater by females (5·3 (se 0·2) μg/d) than by males (3·3 (se 0·2) μg/d). Participants with high income were more likely to be vitamin D supplement users (33·0 %) than those with medium (22·5 %) or low (17·6 %) income. High-income non-Hispanic Whites had the lowest percentage (57 %) not meeting the Estimated Average Requirement for vitamin D. Fortified milk and milk products provided 43·7 % of the dietary vitamin D intake.

Public health efforts should expand the number of vitamin D-fortified foods and encourage the consumption of foods high in vitamin D and use of supplements.

To examine the quantitative relationship between sugar intake and the progressive development of dental caries.

A critical in-depth review of international studies was conducted. Methods included reassessing relevant studies from the most recent systematic review on the relationship between levels of sugars and dental caries. Reanalysis of dose–response relationships between dietary sugars and caries incidence in teeth with different levels of caries susceptibility in children was done using data from Japanese studies conducted by Takeuchi and co-workers.

Global, with emphasis on marked differences in both national sugar intake and fluoride use and preferably where one factor such as sugar intake changed progressively without changes in other factors over a decade or more.

Children aged 6 years or more and adults.

Caries occurred in both resistant and susceptible teeth of children when sugar intakes were only 2–3 % of energy intake, provided that the teeth had been exposed to sugars for >3 years. Despite increased enamel resistance after tooth eruption, there was a progressive linear increase in caries throughout life, explaining the higher rates of caries in adults than in children. Fluoride affects progression of caries development but there still is a pandemic prevalence of caries in populations worldwide.

Previous analyses based on children have misled public health analyses on sugars. The recommendation that sugar intakes should be ≤10 % of energy intake is no longer acceptable. The much greater adult burden of dental caries highlights the need for very low sugar intakes throughout life, e.g. 2–3 % of energy intake, whether or not fluoride intake is optimum.

Most children do not meet the recommended intake of vegetables. Variety was identified as a potential factor to increase children's intake of these foods, as it was shown that variety was effective in improving meal composition in adults. Because younger children are suggested to be more responsive to internal satiation signals than to external food-related cues compared with adults, it is not clear whether variety is effective to improve meal composition in 7- to 10-year-old children.

Experiment.

Children were assigned one of three different fake food buffets containing pasta, chicken, and either one vegetable (carrots or beans) or two vegetables (carrots and beans). The children were asked to serve themselves a meal that they would like to eat for lunch from the given selection.

One hundred children (fifty-two boys; mean age 8·8 (sd 1·1) years).

Children given the two-vegetable choice served themselves significantly more energy from vegetables (mean 64 (sd 51) kJ, 10·9 (sd 9·4) %) compared with children who were offered only either carrots (mean 37 (sd 25) kJ, 5·9 (sd 6·5) %) or beans (mean 38 (sd 34) kJ, 5·6 (sd 6·3) %). The total energy of the meal was not increased, indicating that children chose a more balanced lunch when offered more vegetables.

School-aged children are responsive to food-related cues and variety is effective in increasing their vegetable choice. Serving an assortment of vegetables in school cafeterias might be a simple and effective strategy to improve children's nutrition.

Dietary guidelines are intended to prevent chronic diseases and obesity. The aim of the present study was to develop a diet quality index based on the Spanish Food Pyramid (SFP) and to further explore its association with obesity in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Granada study.

Cross–sectional study. The SFP score considered recommendations given for twelve food groups, and for physical activity and alcohol consumption. Obesity was defined as BMI over 30 kg/m2 and abdominal obesity as waist circumference larger than 102 cm (men) and 88 cm (women). Logistic regression was conducted to estimate odds ratios of obesity by quintiles and by 10-point increment in adherence to the score, controlling for potential confounders.

EPIC-Granada study.

Participants (n 6717) aged 35–69 years (77 % women).

A 10-point increase in adherence to the SFP score was associated with a 14 % (OR=0·86; 95 % CI 0·79, 0·94) lower odds of obesity in men (P interaction by sex=0·02). The odds of abdominal obesity decreased globally by 12 % (OR=0·88; 95 % CI 0·84, 0·93) per 10-point increase in adherence to this score. The effect of higher adherence to the score on abdominal obesity was stronger in physically inactive men and women (ORper 10-point increase=0·79; 95 % CI 0·68, 0·92 and ORper 10-point increase=0·89; 95 % CI 0·84, 0·95, respectively).

These findings support that the Spanish dietary guidelines might be an effective tool for obesity prevention. However, prospective studies investigating this association are warranted.

Fish consumption influences a number of health outcomes. Few studies have directly compared dietary assessment methods to determine the best approach to estimating intakes of fish and its component nutrients, including DHA, and toxicants, including methylmercury. Our objective was to compare three methods of assessing fish intake.

We assessed 30 d fish intake using three approaches: (i) a single question on total fish consumption; (ii) a brief comprehensive FFQ that included four questions about fish; and (iii) a focused FFQ with thirty-six questions about different finfish and shellfish.

Obstetrics practices in Boston, MA, USA.

Fifty-nine pregnant women who consumed ≤2 monthly fish servings.

Estimated intakes of fish, DHA and Hg were lowest with the one-question screener and highest with the thirty-six-item fish questionnaire. Estimated intake of DHA with the thirty-six-item questionnaire was 4·4-fold higher (97 v. 22 mg/d), and intake of Hg was 3·8-fold higher (1·6 v. 0·42 μg/d), compared with the one-question screener. Plasma DHA concentration was correlated with fish intake assessed with the one-question screener (Spearman r = 0·27, P = 0·04), but not with the four-item FFQ (r = 0·08, P = 0·54) or the thirty-six-item fish questionnaire (r = 0·01, P = 0·93). In contrast, blood and hair Hg concentrations were similarly correlated with fish and Hg intakes regardless of the assessment method (r = 0·35 to 0·52).

A longer questionnaire provides no advantage over shorter questionnaires in ranking intakes of fish, DHA and Hg compared with biomarkers, but estimates of absolute intakes can vary by as much as fourfold across methods.

To measure and describe the prevalence and severity of household food insecurity in a remote on-reserve First Nations community using the Household Food Security Survey Module (HFSSM) and to evaluate the perceived relevance of the HFSSM for this population.

Household food security status was determined from the eighteen-item HFSSM following the classifications developed by Health Canada for the Canadian Community Health Survey, Cycle 2·2 Nutrition. One adult from each household in the community was invited to complete the HFSSM and to comment on its relevance as a tool to measure food security for First Nations communities.

Sub-Arctic Ontario, Canada.

Households (n 64).

Seventy per cent of households were food insecure, 17 % severely and 53 % moderately. The prevalence of food insecurity in households with children was 76 %. Among respondents from homes rated as having severe food insecurity, all (100 %) reported worrying that food would run out, times when food didn't last and there wasn't money to buy more, and times when they couldn't afford to eat balanced meals. The majority of respondents felt the HFSSM did not capture an accurate picture of food security for their situation. Aspects missing from the HFSSM included the high cost of market food and the incorporation of traditional food practices.

A high prevalence of household food insecurity was reported in this community. On-reserve remote First Nations communities may be more susceptible to food insecurity than off-reserve Aboriginal populations. Initiatives that promote food security for this vulnerable population are needed.

To validate an FFQ to assess antioxidant intake in overweight postmenopausal women.

A seventy-four-item antioxidant 1-month FFQ was developed based on major antioxidant sources in the American diet. Forty overweight postmenopausal women participated in a 9-month observational study and completed four sets of FFQ and 7 d food record (7dFR) every 3 months. Twelve-hour fasting blood was collected for plasma antioxidant measurement at the first visit.

Connecticut, USA.

Forty overweight postmenopausal women.

Spearman correlation coefficients of 1-month antioxidant intake estimated from the first set of FFQ and 7dFR ranged from 0·34 to 0·87, except for γ-tocopherol. The proportion of participants categorized into the extremely opposite tertiles averaged 7 %. Significant correlations were observed for diet–plasma vitamin C, α-tocopherol and carotenoids (P < 0·05). No time effect was observed on the dietary antioxidant intakes estimated from four 7dFR and four FFQ. Dietary antioxidants estimated from averaged four 7dFR showed moderate to high correlation with those estimated from averaged four FFQ and from each FFQ collected every 3 months. Bland–Altman plots did not show any systematic bias. Averaged misclassifications were below 10 % between these two instruments.

These findings attested a reasonable validity and a good acceptance of this 1-month FFQ in assessing both short-term and long-term diverse antioxidant intakes in these overweight postmenopausal women. The use of this FFQ in associating antioxidant intake with disease risk needs further investigation.

To evaluate the reliability and relative validity of a semi-quantitative FFQ for assessing the habitual intake of multiple nutrients in New Zealand (NZ) adults over the past 12 months.

A 154-item FFQ was developed. After initial pre-testing, reliability was assessed using intra-class correlations. Relative validity was assessed by comparing nutrient intakes derived from the FFQ v. those from an 8 d diet record (8dWDR) collected over 12 months and selected blood biomarkers, using Spearman correlations. Supplementary cross-classification and Bland–Altman analyses were performed to assess validity of the FFQ v. the 8dWDR.

Dunedin, NZ.

One hundred and thirty-two males and females aged 30–59 years who completed all FFQ and 8dWDR and provided a blood sample.

Reliability coefficients ranged from 0·47 for Ca to 0·83 for alcohol, with most values falling between 0·60 and 0·80. The highest validity coefficients for energy-adjusted data were observed for alcohol (0·74), cholesterol (0·65) and β-carotene (0·58), and the lowest for Zn (0·24) and Ca (0·28). For all energy-adjusted nutrients mean percentage correct classification was 77·9 % and gross misclassification was 4·5 %. Results of Bland–Altman analyses showed wide limits of agreement for all micronutrients but high agreement was observed for most macronutrients (99 % for protein, 103 % for total fat). When compared with biomarkers, energy-adjusted coefficients were 0·34 for β-carotene and 0·33 for vitamin C.

The FFQ provides highly repeatable measurements and good validity in ranking individuals’ intake, suggesting that it will be a useful tool to assess nutrient intake of NZ adults in future research.

Urbanization is generally associated with increased CVD risk and accompanying dietary changes. Little is known regarding the association between increased CVD risk and dietary changes using approaches such as diet quality. The relevance of predefined diet quality scores (DQS) in non-Western developing countries has not yet been established.

The association between dietary intakes and CVD risk factors was investigated using two DQS, adapted to the black South African diet. Dietary intake data were collected using a quantitative FFQ. CVD risk was determined by analysing known CVD risk factors.

Urban and rural areas in North West Province, South Africa.

Apparently healthy volunteers from the South African Prospective Urban and Rural Epidemiological (PURE) study population (n 1710).

CVD risk factors were significantly increased in the urban participants, especially women. Urban men and women had significantly higher intakes of both macro- and micronutrients with macronutrient intakes well within the recommended CVD guidelines. While micronutrient intakes were generally higher in the urban groups than in the rural groups, intakes of selected micronutrients were low in both groups. Both DQS indicated improved diet quality in the urban groups and good agreement was shown between the scores, although they seemed to measure different aspects of diet quality.

The apparent paradox between improved diet quality and increased CVD risk in the urban groups can be explained when interpreting the cut-offs used in the scores against the absolute intakes of individual nutrients. Predefined DQS as well as current guidelines for CVD prevention should be interpreted with caution in non-Western developing countries.

The purpose of the present study was to explore the relationship between family meals and parental encouragement of healthy eating overall and by ethnicity.

Family meal frequency was measured with one item asking how many times in the past 7 d all or most of the family ate a meal together, which was then categorized to represent three levels of family meals (≤2 times, 3–6 times and ≥7 times). Parental encouragement of healthy eating assessed how often parents encouraged the student to eat fruits and vegetables, drink water, eat wholegrain bread, eat breakfast and drink low-fat milk (never to always). An overall scale of parental encouragement of healthy eating was created. Mixed-effect regression analyses were run controlling for gender, ethnicity, age and socio-economic status. Moderation by ethnicity was explored.

Middle schools.

Participants included 2895 US eighth grade students participating in the Central Texas CATCH (Coordinated Approach To Child Health) Middle School Project (mean age 13·9 years; 24·5 % White, 52·7 % Hispanic, 13·0 % African-American, 9·8 % Other; 51·6 % female).

Eating more family meals was significantly associated with having parents who encouraged healthy eating behaviours (P for trend <0·001). The number of family meals was positively associated with encouragement of each of the healthy eating behaviours (P for trend <0·0001). There were no differences in the relationships by ethnicity of the students.

Families who eat together are more likely to encourage healthy eating in general. Interventions which promote family meals may include tips for parents to increase discussions about healthy eating.

Current indices of diet quality generally include intakes of specific foods or nutrients. We sought to develop an index that discriminates healthy and unhealthy eating choices for use in large surveys as a short questionnaire and as a measure in existing studies with adequate dietary data.

The Eating Choices Index (ECI) score included four components: (i) consumption of breakfast, (ii) consumption of two portions of fruit per day, (iii) type of milk consumed and (iv) type of bread consumed, each providing a score from 1 to 5. In analysis of 5 d food records, the ECI score was examined in relation to macronutrients, fibre, vitamin C, Fe, Ca and folate using Pearson correlations. Variation with sex, BMI, socio-economic status, marital status, smoking status and physical activity were also investigated.

Medical Research Council National Survey of Health and Development.

Individuals (n 2256) aged 43 years.

The ECI score (mean 12·3 (sd 3·5)) was significantly positively associated with protein, carbohydrate, fibre, vitamin C, Fe, Ca and folate (r = 0·2–0·5; P < 0·001) and significantly negatively associated with fat intake (r = –0·2; P < 0·001); ECI scores were not correlated with total energy intake. Individuals with a lower ECI score were more likely to be men (P < 0·001), overweight or obese (P < 0·001), have lower socio-economic status (P < 0·001), smoke more (P < 0·001) and be less physically active (P < 0·001).

ECI scores correlated with nutrient profiles consistent with a healthy diet. It provides a simple method to rank diet healthiness in large observational studies.

To evaluate the prevalence of vitamin D deficiency and predictors for low vitamin D status in Korean adolescents living between latitudes 33° and 39°N.

A descriptive cross-sectional study.

Korea National Health and Nutrition Examination Survey (KNHANES) 2008–2009.

A total of 1510 healthy adolescents aged 12–18 years (806 male, mean age 14·7 years) participated. Possible predictors for low vitamin D status (log-transformed 25-hydroxyvitamin D (25(OH)D) concentrations) were evaluated.

The prevalence of vitamin D deficiency (25(OH)D<20 ng/ml) was 89·1 % in spring, 53·7 % in summer, 63·9 % in autumn and 90·5 % in winter. Winter season, older age, higher education level reached, being female, being obese, a lack of vitamin D supplementation, lower milk consumption (0–<200 ml/d) and a lack of physical activity were unadjusted predictors (all P < 0·05). Multiple linear regression analysis showed that winter season (P < 0·001), higher education level (P < 0·001) and a lack of vitamin D supplementation (P = 0·012) were independent predictors for low vitamin D status. The modifying effect of season on the association between vitamin D supplement use and vitamin D status was significant (P < 0·001).

Vitamin D deficiency was highly prevalent in Korean adolescents, especially those in higher school grades. Vitamin D supplementation may contribute to maintain a better vitamin D status with lower seasonal variation. Further studies are required to determine optimal vitamin D intakes to maintain sufficient vitamin D status for Korean adolescents.

To describe fruit and vegetable intake of 11-year-old children in ten European countries and compare it with current dietary guidelines.

Cross-sectional survey. Intake was assessed using a previously validated questionnaire containing a pre-coded 24 h recall and an FFQ which were completed in the classroom. Portion sizes were calculated using a standardized protocol.

Surveys were performed in schools regionally selected in eight countries and nationally representative in two countries.

A total of 8158 children from 236 schools across Europe participating in the PRO GREENS project.

The total mean consumption of fruit and vegetables was between 220 and 345 g/d in the ten participating countries. Mean intakes did not reach the WHO population goal of ≥400 g/d in any of the participating countries. Girls had a significantly higher intake of total fruit and vegetables than boys in five of the countries (Sweden, Finland, Iceland, Bulgaria and Slovenia). Mean total fruit intake ranged between 114 and 240 g/d and vegetable intake between 73 and 141 g/d. When using the level ≥400 g/d as a cut-off, only 23·5 % (13·8–37·0 %) of the studied children, depending on country and gender, met the WHO recommendation (fruit juice excluded).

Fruit and vegetable consumption was below recommended levels among the schoolchildren in all countries and vegetable intake was lower than fruit intake. The survey shows that there is a need for promotional activities to improve fruit and vegetable consumption in this age group.

To derive dietary patterns using principal components analysis from separate FFQ completed by mothers and their teenagers and to assess associations with nutrient intakes and sociodemographic variables.

Two distinct FFQ were completed by 13-year-olds and their mothers, with some overlap in the foods covered. A combined data set was obtained.

Avon Longitudinal Study of Parents and Children (ALSPAC), Bristol, UK.

Teenagers (n 5334) with adequate dietary data.

Four patterns were obtained using principal components analysis: a ‘Traditional/health-conscious’ pattern, a ‘Processed’ pattern, a ‘Snacks/sugared drinks’ pattern and a ‘Vegetarian’ pattern. The ‘Traditional/health-conscious’ pattern was the most nutrient-rich, having high positive correlations with many nutrients. The ‘Processed’ and ‘Snacks/sugared drinks’ patterns showed little association with important nutrients but were positively associated with energy, fats and sugars. There were clear gender and sociodemographic differences across the patterns. Lower scores were seen on the ‘Traditional/health conscious’ and ‘Vegetarian’ patterns in males and in those with younger and less educated mothers. Higher scores were seen on the ‘Traditional/health-conscious’ and ‘Vegetarian’ patterns in girls and in those whose mothers had higher levels of education.

It is important to establish healthy eating patterns by the teenage years. However, this is a time when it is difficult to accurately establish dietary intake from a single source, since teenagers consume increasing amounts of foods outside the home. Further dietary pattern studies should focus on teenagers and the source of dietary data collection merits consideration.

To assess sociodemographic correlates of micronutrient intakes from food and dietary supplements in an urban, ethnically diverse sample of pregnant women in the USA.

Cross-sectional analyses of data collected using a validated semi-quantitative FFQ. Associations between racial, ethnic and sociodemographic factors and micronutrient intakes were examined using logistic regression controlling for pre-pregnancy BMI, maternal age and smoking status.

Prenatal clinics, Boston, MA, USA.

Analyses included pregnant women (n 274) in the PRogramming of Intergenerational Stress Mechanisms (PRISM) study, an urban longitudinal cohort designed to examine how stress influences respiratory health in children when controlling for other environmental exposures (chemical stressors, nutrition).

High frequencies of vitamin E (52 %), Mg (38 %), Fe (57 %) and vitamin D (77 %) inadequacies as well as suboptimal intakes of choline (95 %) and K (99 %) were observed. Factors associated with multiple antioxidant inadequacies included being Hispanic or African American, lower education and self-reported economic-related food insecurity. Hispanics had a higher prevalence of multiple methyl-nutrient inadequacies compared with African Americans; both had suboptimal betaine intakes and higher odds for vitamin B6 and Fe inadequacies compared with Caucasians. Nearly all women (98 %) reported Na intakes above the tolerable upper limit; excessive intakes of Mg (35 %), folate (37 %) and niacin (38 %) were also observed. Women reporting excessive intakes of these nutrients were more likely Caucasian or Hispanic, more highly educated, US-born and did not report food insecurity.

Racial/ethnic and other sociodemographic factors should be considered when tailoring periconceptional dietary interventions for urban ethnic women in the USA.

To evaluate the dietary quality of Mexican adults’ diet, we constructed three dietary quality indices: a cardioprotective index (CPI), a micronutrient adequacy index (MAI) and a dietary diversity index (DDI).

Data were derived from the 2006 National Health and Nutrition Survey, which is a national survey representative of the Mexican population with a stratified, multistage, probabilistic sample design. Dietary intake was assessed from an FFQ with 101 different foods and daily nutrient intakes were computed. The CPI evaluated compliance with seven WHO recommendations for the prevention of CVD, the MAI evaluated the intake of six micronutrients based on the estimated average requirements from the US Institute of Medicine and the DDI was constructed based on the consumption of thirty different food groups.

Mexico.

Mexican adults aged 19–59 years old.

We evaluated the diet of 15 675 males and females. Adjusted means and adjusted proportions by age and sex were computed to predict adherence to dietary recommendations. Rural inhabitants, those living in the South and those from the lowest socio-economic status reported a significantly higher CPI (4·5 (se 0·08), 4·3 (se 0·08) and 4·2 (se 0·09), respectively; P < 0·05), but a significantly lower MAI and DDI, compared with urban inhabitants, those from the North and those of upper socio-economic status (P < 0·05).

The constructed diet quality indices identify nutrients and foods whose recommended intakes are not adequately consumed by the population. Given the epidemiological and nutritional transition that Mexico is experiencing, the CPI is the most relevant index and its components should be considered in Mexican dietary guidelines as well as in any food and nutrition programmes developed.

To assess parents’ perspectives of their home environments to establish the validity of scores from the Behavior and Attitudes Questionnaire for Healthy Habits (BAQ-HH).

In the present descriptive study, we surveyed a cross-sectional sample of parents of pre-school children. Questionnaire items developed in an iterative process with community-based programming addressed parents’ knowledge/awareness, attitudes/concerns and behaviours about healthy foods and physical activity habits with 6-point rating scales. Exploratory and confirmatory factor analyses were used to psychometrically evaluate scores from the scales.

English and Spanish versions of the BAQ-HH were administered at parent–teacher conferences for pre-school children at ten Head Start centres across a five-county agency in autumn 2010.

From 672 families with pre-school children, 532 parents provided responses to the BAQ-HH (79 % response rate). The majority was female (83 %), Hispanic (66 %) or white (16 %), and ages ranged from 20 to 39 years (85 %).

Exploratory and confirmatory analyses revealed a knowledge scale (seven items), an attitude scale (four items) and three behaviour subscales (three items each). Correlations were identified between parents’ perceptions of home activities and reports of children's habits. Differences were identified by gender and ethnicity groupings.

As a first step in psychometric testing, the dimensionality of each of the three scales (Knowledge, Attitudes and Behaviours) was identified and scale scores were related to other indicators of child behaviours and parents’ demographic characteristics. This questionnaire offers a method to measure parents’ views to inform planning and monitoring of obesity-prevention education programmes.

We investigated factors accounting for the consistently higher levels of household food insecurity reported by women in Canada.

Two cycles of the Canadian Community Health Survey for the years 2005/2006 and 2007/2008 were pooled to examine the association between household food insecurity, measured using the Household Food Security Survey Module and other metrics, and respondent sex. We stratified households as married/cohabiting (in which case, the household respondent was chosen randomly) or non-married (single/widowed/separated/divorced) and adjusted for differences in household characteristics, including the presence of children.

Canada.

Analysis was restricted to households dependent on employment/self-employment and whose reported annual household income was below $CAN 100 000. Exclusions included respondents less than 18 years of age, any welfare receipt, and missing food insecurity, marital status, income source and amount, or household composition data.

For non-married households, increased food insecurity in female- v. male-led households was accounted for by significant differences in household socio-economic characteristics. In contrast, in married/cohabiting households with or without children, higher food insecurity rates were reported when the respondent was female and neither respondent characteristics nor socio-economic factors accounted for the differences.

Higher rates of food insecurity in non-married households in Canada are largely attributable to women's socio-economic disadvantage. In married households, women appear to report higher levels of food insecurity than men. These findings suggest a possible bias in the measurement of population-level household food insecurity in surveys that do not account for the sex of the respondent in married/cohabiting households.

To explore multiple methods of calculating diet diversity scores (DDS) to maximize associations with predicted dietary micronutrient adequacy among schoolchildren in rural Kenya.

Up to three 24 h recall interviews were administered for each child for a total of 1544 d of intake from all schoolchildren. Daily amounts of food consumed were assigned to one of eight food groups. Five DDS were developed based on various minimum intake amounts from each food group: (i) 1 g; (ii) 15 g; (iii) a variable minimum based on the content of a target nutrient for each group; (iv) the median intake level for each group; and (v) the 90th percentile intake level for each group. A diet was assigned 1 point towards the daily DDS if the food group intake was above the defined minimum level. Five scores were calculated for each child, and bivariate longitudinal random-effects models were used to assess the correlation between each DDS and the mean probability of adequacy for fourteen nutrients.

Embu District, Kenya.

Schoolchildren (n 529), mean age 7·00 (sd 1·41) years.

Only DDS based on a 15 g minimum and DDS based on nutrient content were significantly associated with mean probability of adequacy after adjusting for energy intake (0·21 and 0·41, respectively).

A DDS using minimum intakes based on nutrients contributed by a food group best predicted nutrient adequacy in this population. These analyses contribute to the continued search for simpler and more valid dietary quality indicators among low-income nations.