Exploration of the Principal Component Analysis (PCA) Approach in Synthesizing the Diet Quality of the Malaysian Population.

(1) Background: One of the most main dietary assessments is through a posteriori application. Although extensive research has incorporated dietary assessment of a population through a posteriori application, this study is the first to examine the Malaysian population and use an a posteriori method and principal component analysis (PCA) to assess the dietary patterns of the Malaysian population. The correlation between all dietary patterns derived via PCA and selected nutrient intake were determined in this sample of study; (2) Methods: A total of 3063 respondents (18 to 59 years old) covering Peninsular Malaysia, Sabah, and Sarawak, participated in this study. PCA was applied on the food frequency questionnaire collected from the respondents, and descriptive statistics and PCA were performed using SPSS version 21; (3) Results: Six patterns were identified: "traditional", " prudent", " modern", "western", "Chinese", and "combination" diets. All together, these six patterns were able to explain 45.9% of the total variability. Few components derived from the factor loadings showed positive association with several nutrient markers. The traditional dietary pattern showed a moderate, positive correlation with total protein and total sugar intake, there was a significant moderate correlation between the prudent dietary pattern and dietary fibre, and there was a moderate positive association between the Chinese dietary pattern and total energy; and (4) Conclusions: The exploration of the PCA approach above may provide justification for assessment of dietary patterns rather than reliance on single nutrients or foods to identify potential connections to overall nutritional wellbeing as well as to explore the diet-disease relationship. However, study of pattern analysis must be conducted among the Malaysian population to produce validity and reproducibility for this dietary approach in light of the numerous methodological issues that arise when performing PCA.

Effectiveness of a multi-disciplinary family-based programme for treating childhood obesity (the Family Project).

BACKGROUND/OBJECTIVES: To determine if a multi-component family focused education package is more effective than a waiting list control group in treating overweight and obese children. SUBJECTS/METHODS: A 2-year randomised controlled trial; 65 overweight and obese children aged 6-14 years were allocated to active intervention in either the first or second year, with body composition monitoring alone in the control period. Anthropometric measurements were undertaken at six monthly intervals and a 7-day food and activity diary were issued. RESULTS: Over the 2 years of the study body mass index (BMI) SDS (z score) fell significantly in the intervention/control (I/C) group, but not in the control/intervention (C/I) group. The difference between groups was 0.3, which was borderline significant (95% confidence interval (95% CI) -0.62 to 0.02, P=0.06) before adjusting for potential confounding factors. Thirty-three percent of the I/C group and 12% of the C/I group achieved the target reduction of 0.5 BMI SDS. The I/C group had a significantly greater reduction in the percentage with a BMI above the 99.6th centile at 24 months (P=0.04) and gained 5.7 kg less over the time of the study. There were no significant differences between groups for mean percentage attendance at physical activity sessions (I/C group=24.1%, 95% CI, 15.4-32.9; C/I group=31.7%, 95% CI, 22.4-41.1, P=0.229). CONCLUSIONS: Children given active intervention followed by body composition monitoring alone reduced their BMI SDS, and fewer children were classified as grossly overweight by the end of the study. If these findings are true, there are important implications for the provision of services managing overweight in the community.

Changing health behaviour of young women from disadvantaged backgrounds: evidence from systematic reviews.

Observational evidence suggests that improving the diets of women of child-bearing age from disadvantaged backgrounds might be an important component of public health strategies aimed at reducing the burden of chronic disease in their offspring. The development of an intervention to improve the nutrition of young women needs to be informed by a systematic collation of evidence. Such a systematic collation of evidence from systematic reviews of interventions directed at changing health behaviours including diet, breast-feeding, physical activity and smoking has been conducted. Of 1847 potentially-relevant abstracts, fourteen systematic reviews met inclusion criteria. Four aspects of intervention design were identified that were effective at changing one or more of the health behaviours considered in the present review: the use of an educational component; provision of continued support after the initial intervention; family involvement; social support from peers or lay health workers. The findings of the present review suggest that interventions to change the health behaviour of women of child-bearing age from disadvantaged backgrounds will require an educational approach and should provide continued support after the initial intervention. Family involvement and social support from peers may also be important features of interventions that aim to improve diet.

Constraints on food choices of women in the UK with lower educational attainment.

OBJECTIVE: Women of lower educational attainment have less balanced and varied diets than women of higher educational attainment. The diets of women are vital to the long-term health of their offspring. The present study aimed to identify factors that influence the food choices of women with lower educational attainment and how women could be helped to improve those choices. DESIGN: We conducted eight focus group discussions with women of lower educational attainment to identify these factors. We contrasted the results of these discussions with those from three focus group discussions with women of higher educational attainment. SETTING: Southampton, UK. SUBJECTS: Forty-two white Caucasian women of lower educational attainment and fourteen of higher educational attainment aged 18 to 44 years. RESULTS: The dominant theme in discussions with women of lower educational attainment was their sense that they lacked control over food choices for themselves and their families. Partners and children exerted a high degree of control over which foods were bought and prepared. Women's perceptions of the cost of healthy food, the need to avoid waste, being trapped at home surrounded by opportunities to snack, and having limited skill and experience with food, all contributed to their sense they lacked control over their own and their family's food choices. CONCLUSIONS: An intervention to improve the food choices of women with lower educational attainment needs to increase their sense of control over their diet and the foods they buy. This might include increasing their skills in food preparation.

Cardiovascular disease risk factors and socio-economic position of Africans in transition: the THUSA study.

In many developing countries with advanced stages of the nutrition transition, the burden of coronary artery disease (CAD) has shifted from the rich to the poor. In South Africa, it is mainly the African population that is experiencing rapid urbanisation and the nutrition transition. It is not clear where the burden of CAD lies in this population group. We tried to answer this question by comparing CAD risk factors within African groups of different socio-economic positions (characterised by total household income and education level) that participated in the THUSA study from 1996 to 1998. The THUSA study was a cross-sectional population- based epidemiological study that examined the influence of urbanisation and related changes in lifestyle and eating patterns on health and disease risk. A total of 1 854 apparently healthy African volunteers were recruited from 37 randomly chosen sites in rural and urban areas of the North-West Province. The results indicated that although the group with the highest socio-economic position had significantly lower serum glucose levels, systolic blood pressures, higher micronutrient intakes and fewer smokers, their sustained increases in total and saturated fat intakes and higher serum total and LDL cholesterol levels, as well as increased body mass indices in men suggested that at that point in time and possibly in the foreseeable future, the burden of CAD will be carried by those Africans with higher socio-economic positions.

Rapid assessment of maternal activity among rural Indian mothers (Pune Maternal Nutrition Study).

OBJECTIVE: To develop an activity questionnaire for objective and rapid assessment of maternal habitual physical activity with the aim of describing its relationship with birth size. DESIGN: Prospective observational study. SETTING: Six villages near Pune, Maharashtra, India. SUBJECTS: Seven hundred and ninety-seven rural Indian mothers were studied after excluding abortions and termination of pregnancies, foetal anomalies, etc. METHOD: Principal components analysis (PCA) was used to identify a few leading questions from an elaborate questionnaire involving 36 questions related to 14 typical maternal activities. RESULTS: On the basis of high loading (>0.8), PCA identified three activities, i.e. farming, fetching water and washing clothes. Questions and sub-questions related to these activities only were considered for defining an activity score for rapid assessment. Validation of this new activity score, with the score based on the elaborate questionnaire, showed 70% sensitivity as well as specificity. New activity scores showed strong inverse relationships similar to those observed using the elaborate questionnaire, at early as well as late gestation, with neonatal head circumference (P=0.001 and 0.055) and mid-arm circumference (P=0.02 and 0.03). CONCLUSIONS: Simple questionnaires, based on leading activities identified by PCA, can be as informative as longer and detailed questionnaires. This method has potential for adaptation, especially in rural communities in developing countries.

Integrated nutrition science: from theory to practice in South Africa.

OBJECTIVE: To describe an integrated, holistic conceptual framework and research paradigm for a better understanding of the nutrition transition in middle- and low-income countries. MOTIVATION: Current inability effectively to prevent the increasing burden related to changes in food consumption patterns and other health behaviours of populations in transition motivates a new approach for nutrition research and practice. In this proposed approach, broader and integrated dimensions of science and practice may be applied for a better understanding of this complex phenomenon. RESULT: Examples from our own studies are given and quoted to illustrate how results from transdisciplinary studies were used to design an integrated, holistic programme to improve quality of life of people infected with HIV. CONCLUSION: Based on these experiences it is argued that the more holistic and integrated approach should and could lead to more effective and sustainable interventions to prevent the adverse health consequences of the nutrition transition. At the same time such an approach will contribute to efforts to conserve the environment and also human, living and natural resources.

Food-frequency questionnaires: a review of their design, validation and utilisation.

A review of the literature concerning the design, utilisation and validation of food-frequency questionnaires (FFQ) has been carried out using a semi-systematic approach to obtaining, reviewing and extracting data from articles. Databases were searched from 1980 to 1999. The present review identified 227 validation (from 1980 to September 1999) and 164 utilisation (for 1998 only) studies. A number of design issues have been evaluated through the present review. These include: the need to consider how portion sizes have been described, self-defined giving higher mean correlations; how an FFQ was administered, interviewer-administered giving higher mean correlations for some nutrients; how many items to include on an FFQ, those with the largest number of items having higher correlations. Validation techniques were described. Most validation studies involved comparing an FFQ against another dietary assessment method; only 19 % compared an FFQ to a biomarker. Measurement differences were most commonly assessed by correlation coefficients as opposed to other more appropriate methods. Mean correlation coefficients were highest for Ca and fat, and lowest for vitamin A and vegetables. The utilisation studies showed that FFQ were most commonly used in cross-sectional surveys, with ninety-three of the FFQ being designed to be disease-specific. The present review results were presented to a group of experts and a consensus arrived at concerning the development, validation and use of FFQ. Recommendations derived from the consensus arising from the literature review are presented as an appendix to the present paper.

Maternal activity in relation to birth size in rural India. The Pune Maternal Nutrition Study.

OBJECTIVE: To describe the relationship of the mother's physical activity to the birth size of her baby in a rural Indian population. DESIGN: : Prospective observational study. SETTING: Six villages near Pune, Maharashtra, India. SUBJECTS: : A total of 797 women were studied after excluding abortions and termination of pregnancies (112), foetal anomalies (8), multiple pregnancies (3), incomplete pre-pregnancy anthropometry (14) and pregnancies detected later than 21 weeks of gestation (168). METHOD: An activity questionnaire was developed after focus group discussions and incorporated community-specific activities. It was validated against an observer-maintained diary. Activity scores were derived using published data on energy costs to weight the contributions of various activities. It was then administered to assess physical activity at 18 (+/-2) and 28 (+/-2) weeks of gestation. OUTCOME MEASURES: Birth outcome, maternal weight gain and neonatal anthropometry. RESULTS: The activity questionnaire was used to classify women into light, moderate and heavy activity categories. Maternal activity did not influence the incidence of prematurity or stillbirth, or the duration of gestation. It was inversely related to maternal weight gain up to 28 weeks of gestation (P=0.002). Higher maternal activity in early, as well as mid gestation, was associated with lower mean birth weight (P=0.05 and 0.02, respectively ), and smaller neonatal head circumference (P=0.005 and 0.009) and mid-arm circumference (P=0.03 and 0.01) after adjusting for the effect of major confounding factors. CONCLUSIONS: The Findings suggest that excessive maternal activity during pregnancy is associated with smaller foetal size in rural India, The approach described for developing an activity questionnaire has potential for adoption in other settings.

Prevalence of risk of undernutrition is associated with poor health status in older people in the UK.

OBJECTIVES: To establish the prevalence of the risk of undernutrition, using criteria similar to those used by the Malnutrition Advisory Group (MAG), in people aged 65 y and over, and to identify relationships between risk of undernutrition and health and demographic characteristics. DESIGN: A cross-sectional nationally representative sample of free-living and institutionalized older people in the UK (65 y of age and over). Secondary analysis of the National Diet and Nutrition Survey based on 1368 people aged 65 y and over. RESULTS: About 14% (21% in those living in institutions) were at medium or high risk of undernutrition based on a composite measure of low body mass index and recent reported weight loss. Having a long-standing illness was associated with a statistically significantly increased risk of undernutrition (odds ratio: men 2.34, 95% CI 1.20-4.58; women 2.98; 1.58-5.62). The risk of undernutrition increased: in women reporting bad or very bad health status; in men living in northern England and Scotland; for those aged 85 y and older; for those hospitalized in the last year, and those living in an institution. Lower consumption of energy, meat products or fruit and vegetables and lower blood measures of zinc, vitamins A, D, E and C were associated with statistically significantly increased risk of undernutrition. CONCLUSIONS: A substantial proportion of the older population of the UK is at risk of undernutrition. High-risk subjects are more likely to have poorer health status. It is unlikely that the individuals at high risk are being detected currently, and therefore effective care is not being provided, either in the community or in institutions.

Persistence of lower birth weight in second generation South Asian babies born in the United Kingdom.

OBJECTIVE: To assess differences in birth weight between all first and second generation South Asian babies born in Southampton, and trends since 1957. DESIGN: Retrospective, cohort study. SETTING: Birth records for babies born in Southampton from 1957 to 1996 were searched to identify all babies born of South Asian origin (including from the Indian subcontinent, East Africa, and elsewhere). MAIN OUTCOME MEASURES: All information recorded in the birth record about the mother and baby was extracted. RESULTS: 2395 full term (>37 weeks; mean birth weight 3110; 95%CI 3092 to 3129) singleton births were identified. Detailed analysis was restricted to mothers either born in the Indian subcontinent (India, Pakistan, or Bangladesh (1435)) or United Kingdom (283). Mean birth weight and % low birth weight (<2500 g) were 3133 g (95%CI 3108 to 3157) and 7.5%, for first generation babies and 3046 g (2992 to 3099) and 11.7% for second generation babies. There was no trend over time to increased average birth weight in either first or second generation babies. Adjusting for other factors that were statistically significantly related to birth weight (gender, gestational age, mother's age, maternal weight at 15 weeks, parity, and mother's ethnic group) did not alter the trends. CONCLUSIONS: For that group in the UK who derive from the Indian subcontinent, average birth weight is significantly less than the national average. There has not been any increase in the average birth weight over the past 40 years, and the birth weight of babies of women who were born in the UK are no greater. The persistence of lower than desirable birth weight may result long term in higher than average rates of diabetes and heart disease in these groups.

Dietary antioxidants and asthma in adults: population-based case-control study.

A protective role for dietary antioxidants in asthma has been proposed. However, epidemiological evidence to implicate antioxidant vitamins is weak, and data on the role of flavonoid-rich foods and antioxidant trace elements are lacking. We carried out a population-based case-control study in South London, UK, to investigate whether asthma is less common and less severe in adults who consume more dietary antioxidants. Participants were aged 16-50 yr and registered with 40 general practices. Asthma was defined by positive responses to a standard screening questionnaire in 1996, and complete information about usual diet was obtained by food frequency questionnaire from 607 cases and 864 controls in 1997. After controlling for potential confounding factors and total energy intake, apple consumption was negatively associated with asthma (odds ratio [OR] per increase in frequency group 0.89 [95% confidence interval [CI]: 0.82 to 0.97]; p = 0.006). Intake of selenium was also negatively associated with asthma (OR per quintile increase 0.84 [0.75 to 0.94]; p = 0.002). Red wine intake was negatively associated with asthma severity. The associations between apple and red wine consumption and asthma may indicate a protective effect of flavonoids. The findings for dietary selenium could have implications for health policy in Britain where intake has been declining.

Intake of micronutrient-rich foods in rural Indian mothers is associated with the size of their babies at birth: Pune Maternal Nutrition Study.

One third of the Indian babies are of low birth weight (<2.5 kg), and this is attributed to maternal undernutrition. We therefore examined the relationship between maternal nutrition and birth size in a prospective study of 797 rural Indian women, focusing on macronutrient intakes, dietary quality and micronutrient status. Maternal intakes (24-h recall and food frequency questionnaire) and erythrocyte folate, serum ferritin and vitamin C concentrations were measured at 18 +/- 2 and 28 +/- 2 wk gestation. Mothers were short (151.9 +/- 5.1 cm) and underweight (41.7 +/- 5.1 kg) and had low energy and protein intakes at 18 wk (7.4 +/- 2.1 MJ and 45.4 +/- 14.1 g) and 28 wk (7.0 +/- 2.0 MJ and 43.5 +/- 13.5 g) of gestation. Mean birth weight and length of term babies were also low (2665 +/- 358 g and 47.8 +/- 2.0 cm, respectively). Energy and protein intakes were not associated with birth size, but higher fat intake at wk 18 was associated with neonatal length (P < 0.001), birth weight (P < 0.05) and triceps skinfold thickness (P < 0.05) when adjusted for sex, parity and gestation. However, birth size was strongly associated with the consumption of milk at wk 18 (P < 0.05) and of green leafy vegetables (P < 0.001) and fruits (P < 0.01) at wk 28 of gestation even after adjustment for potentially confounding variables. Erythrocyte folate at 28 wk gestation was positively associated with birth weight (P < 0.001). The lack of association between size at birth and maternal energy and protein intake but strong associations with folate status and with intakes of foods rich in micronutrients suggest that micronutrients may be important limiting factors for fetal growth in this undernourished community.

The incidence and distribution of Legg-Calvé-Perthes' disease in Liverpool, 1982-95.

AIMS: To determine the incidence and distribution of Legg-Calvé-Perthes' disease in Liverpool, in the period 1982-95. METHODS: Examination of information in a register, analysing the patients' addresses by indices of deprivation. RESULTS: A total of 122 white children were diagnosed as having Perthes' disease during the study, whereas black and minority groups form 5.8% of the population. The incidence rate in inner Liverpool had decreased to 10.5 in the period 1990-95. Simple Spearman correlations revealed an association between the disease incidence in electoral wards and deprivation. Regression analysis showed that for the period 1990-95 the most powerful effects on incidence were increases in ward deprivation since 1976, the percentage free school meals in 1986, the ward Health Index in 1981, and the percentage low birth weight in 1981. CONCLUSIONS: We suggest that environmental influences may come into play some years before a child presents with pain in the hip. There may be a genetic predisposition to the disease.

An epidemiological study of hypertension and its determinants in a population in transition: the THUSA study.

BACKGROUND: Many black persons in South Africa have been subjected to urbanisation and urbanisation has led to a significant increase in diseases of lifestyle. The determinants of hypertension in a population in transition have not been well-defined and there is a pressing need for observational epidemiological studies as well as randomised-controlled trials in populations from Africa. The aim of this study was to investigate the association between blood pressure and factors known to contribute to hypertension. METHODS: The study sample consisted mainly of Setswana speaking people, divided into different levels (strata) of urbanisation, namely stratum 1 (rural) to stratum 5 (urbanized). A total of 1821 black subjects, which included 1040 woman, were recruited and randomly selected from 37 sites from the four geographical quarters of the North West Province of South Africa. The following questionnaires were used: demographic, anthropometric, quantitative food frequency, physical activity and scales to measure psychosocial variables. Biochemical analysis (standardised methods) were done on the serum and plasma of the subjects and the blood pressure was measured with a sphygmomanometer. RESULTS: Of the total sample, 22.8% of the subjects had systolic and 20.7% diastolic blood pressures above 140/90 mm Hg. Males and females from stratum 3 showed the highest rate of hypertension (32.9% systolic and 25.1% diastolic) and stratum 5 the lowest. Blood pressure correlated positively with age, level of urbanisation, WHR (waist:hip ratio) and smoking. In the woman the diastolic blood pressure correlated the best with body mass index (BMI), serum triglycerides, total serum cholesterol, low-density lipoprotein (LDL) cholesterol and s-GGT. Coping strategies, experience of social support, cultural aspects and affect balance are related to blood pressure, especially in the case of women. CONCLUSIONS: It seems that factors associated with urbanisation are related to the manifestation of hypertension in black people of the North West Province, given the highest mean blood pressure in people living in informal settlements, where most newcomers to the urban areas live. Journal of Human Hypertension (2000) 14, 779-787

Interaction between physical activity and diet: implications for blood pressure management in primary care.

OBJECTIVES: This review summarises the evidence for the effectiveness of primary care based interventions of diet and physical activity aimed at reducing blood pressure. BACKGROUND: As blood pressure rises so does the risk of heart disease and stroke. There is a large literature on the effects on blood pressure of changing various aspects of diet either as single nutrient interventions, patterns of food consumption, or the addition of dietary supplements (potassium, magnesium, and fish oil). Controlled trials have been undertaken to assess the relative benefits of lifestyle changes in activity (walking etc) compared with more structured exercise programmes. There is sufficient evidence to suggest that changes in diet and activity under controlled conditions can reduce blood pressure and delay or reduce the need for drug treatment in subjects who are hypertensive. What is less clear is how to achieve these lifestyle changes in the primary care setting as part of routine clinical management. RESULTS: There have been a number of systematic reviews undertaken to evaluate the evidence. The interventions were either delivered alone or in combination with other advice, either with or without aids, and by various practice nurse staff (GP, nurse, dietitian); follow-up varied from 3 months to a year. Compliance with the advice was generally not measured. The variability in the quality of the studies and interventions made it difficult to draw conclusions: any effects seen tended to be small. CONCLUSIONS: Equivocal results (non statistically significant reductions) should not be considered as proof of no effect without careful consideration of the effects of chance, bias and confounding, and without better measures of compliance. There is little debate that, under investigator control, blood pressure can be reduced by changes in diet and activity. Changes to a 'healthy' diet (low in total and saturated fat; energy intake balanced with expenditure to maintain or achieve optimal body weight; low in salt: high in fruits, vegetables, legumes; and whole grains) and increases in modest levels of physical activity (walking etc) would be expected to reduce blood pressure by between two and four mmHg. A shift in the population distribution of this order would be expected to have a substantial impact on population mortality patterns and could be achieved cost-effectively in primary care.

The Health Education Authority's health and lifestyle survey 1993: who are the low fruit and vegetable consumers?

STUDY OBJECTIVE: Firstly, to determine the demographic and behavioural characteristics of low fruit and vegetable consumers. Secondly, to investigate whether knowledge and attitudes are barriers to consumption of fruit and vegetables. DESIGN: Cross sectional survey: an interviewer administrated questionnaire was used to assess the demographic, knowledge, attitude, and behavioural characteristics of the respondents. SETTING: England. PARTICIPANTS: Random sample of 5553 men and women aged between 16 and 74 years. Response rate 70%. MAIN RESULTS: The main demographic characteristics of the respondents identified as low consumers of fruit and vegetables (less than daily consumption of either fruit or vegetables) were age, sex, and smoking status. The adjusted odds ratios were 2.59 for those aged 16-24 years compared with those aged 45-74 years, 2.17 for men compared with women, and 1.77 for current smokers compared with never smokers. The most important knowledge and attitude statements after adjusting for the demographic variables were disagreeing with the statement "healthy foods are enjoyable" (odds ratio 1.90) and agreeing with the statement "I don't really care what I eat" (odds ratio 1.76). The impact of knowledge seemed less important than attitudes about a healthy diet in characterising a low fruit and vegetable consumer. CONCLUSIONS: These findings are relevant to future strategies for improving intake of fruit and vegetables, but demonstrate the complexity of interventions required, and the dangers inherent in assuming simplistic relations between psychosocial factors and behaviour.

Relationship between attitudes to health, body weight and physical activity and level of physical activity in a nationally representative sample in the European Union.

OBJECTIVE: To explore the factors that influence attitudes and beliefs about the effects of body weight and physical activity on health. DESIGN: Cross-sectional survey using a face-to-face interview-assisted questionnaire. SETTING: The survey was conducted between March and April 1997 in the 15 member states of the EU. SUBJECTS: Approximately 1000 adults aged 15 years plus from each country were selected by quota-controlled sampling; the total sample was of 15,239 persons (7162 males and 8077 females). Data were weighted by population size for each country and by age, gender and regional distribution within countries. RESULTS: Overall 27% of men and 35% of women reported not engaging in any form of recreational activity; rates were highest in those with a primary-level education (37% men, 43% women) compared with tertiary-level education (20% men, 25% women). Recreational activity levels declined with age and tended to be lowest in those who were more physically active at work. Eighteen per cent of respondents believed that physical activity was one of the two greatest influences on health; 13% mentioned body weight, 38% mentioned food, 41% mentioned smoking and 33% mentioned stress. Logistic regression was used to assess for the effects of gender, age, educational level, body mass index (BMI), smoking status, activity level and country on beliefs that body weight and physical activity influenced health. Primary-educated, overweight, ex- and non-smoking women were most likely to mention body weight as an influence; young tertiary-educated, thinner, non-smoking and active males were most likely to mention physical activity. There was wide variation across Europe in reported behaviour and beliefs, which persisted after taking account of all of the other factors in the regression model. CONCLUSIONS: Relatively few people in some countries believed body weight and physical activity were important influences on health. Many people mentioned stress and it might be possible to use this interest to motivate changes in physical activity to relieve stress. It may be important to take a more integrated approach to activity patterns that consider the role of work and recreation. Although social and demographic factors were important, after adjusting for these factors there was still wide between-country variation in reported attitudes, beliefs and behaviour. This will need to be taken into account in any activity promotion campaigns.